{"meta":{"query_hash":"0289893b6e2e","filters":{"venue":"The International Journal of Biostatistics"},"cohort_total":68,"direct_labels_cover":0,"predictions_cover":68,"exported":68,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/0289893b6e2e","api":"https://metacan.xera.ac/api/v1/cohort?venue=The+International+Journal+of+Biostatistics"},"results":[{"id":"W1729931698","doi":"10.1515/ijb-2016-0004","title":"Multiple Comparisons Using Composite Likelihood in Clustered Data","year":2016,"lang":"en","type":"preprint","venue":"The International Journal of Biostatistics","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Multivariate statistics; Quasi-maximum likelihood; Statistics; Monte Carlo method; Computer science; Gaussian; Mathematics; Algorithm; Data mining; Econometrics; Expectation–maximization algorithm; Maximum likelihood","score_opus":0.683109364844231,"score_gpt":0.5842110733642264,"score_spread":0.09889829148000462,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1729931698","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.013832651,0.0001585629,0.9670631,0.003353422,0.0073250397,0.00039371278,0.007702887,0.000016987296,0.0001535843],"genre_scores_gemma":[0.1689982,0.00015300946,0.8289876,0.00025679704,0.0015086511,0.0000021943135,0.000028645461,0.000050230494,0.000014685202],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99396765,0.0011331004,0.0028317503,0.000381004,0.0013663548,0.0003201285],"domain_scores_gemma":[0.93999857,0.054774296,0.002995412,0.001084937,0.0010014239,0.00014537564],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0058995956,0.0003384636,0.0010793038,0.000267493,0.000055370547,0.00019456251,0.0046634264,0.00027673534,0.00008921746],"category_scores_gemma":[0.06606848,0.00021541031,0.00018617438,0.00009295625,0.00030381983,0.00009898094,0.003546551,0.0014221886,0.000011087861],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.014806869,0.008761457,0.05696252,0.002059965,0.018001698,0.0060235383,0.0032860585,0.00358985,0.014686959,0.21382248,0.43659326,0.22140533],"study_design_scores_gemma":[0.0022747752,0.00006354955,0.0014933135,0.0019340119,0.00036462027,0.00014855122,0.00007520342,0.038408656,0.00033709925,0.9538571,0.0007386048,0.00030451364],"about_ca_topic_score_codex":0.00006540492,"about_ca_topic_score_gemma":0.00008364357,"teacher_disagreement_score":0.74003464,"about_ca_system_score_codex":0.0003466513,"about_ca_system_score_gemma":0.00049556326,"threshold_uncertainty_score":0.94179845},"labels":[],"label_agreement":null},{"id":"W1965231589","doi":"10.2202/1557-4679.1216","title":"Estimation of Modified Concordance Ratio in Sib-Pairs: Effect of Consanguinity on the Risk of Congenital Heart Diseases","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"","keywords":"Consanguinity; Concordance; Confidence interval; Statistics; Demography; Population; Sample (material); Sample size determination; Inference; Estimation; Econometrics; Medicine; Genetics; Mathematics; Biology; Computer science; Environmental health","score_opus":0.008290049680790001,"score_gpt":0.2870186780345715,"score_spread":0.2787286283537815,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1965231589","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99288946,0.00009344024,0.00571433,0.00042723076,0.00030675682,0.000102226506,0.00040036687,5.894902e-7,0.00006558607],"genre_scores_gemma":[0.9985244,0.00007004522,0.0012200228,0.00006917169,0.000076044234,0.0000022651186,0.00002393799,0.0000044943517,0.000009570285],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9989149,0.00026335916,0.0004993869,0.000061358565,0.00019700425,0.00006399872],"domain_scores_gemma":[0.99744725,0.0012335507,0.0008049832,0.00013580117,0.0003570018,0.000021412601],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00089317287,0.00007142091,0.00019530581,0.00004422454,0.000021349262,0.0000047711446,0.00026065327,0.00005343574,0.000014158648],"category_scores_gemma":[0.0049982397,0.000042135627,0.000088249784,0.000041472293,0.00023506595,0.0000030671265,0.000036263547,0.00014346122,5.9391726e-7],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0023577898,0.0004633428,0.46850634,0.00004861263,0.00111177,0.0000064442565,0.00039095688,0.036539286,0.46598256,0.009521392,0.0060601505,0.009011385],"study_design_scores_gemma":[0.002257419,0.0021329804,0.6311197,0.0000823178,0.000216943,0.000042239917,0.00020206728,0.022344084,0.33514366,0.006126151,0.0001835061,0.00014886964],"about_ca_topic_score_codex":0.00007210214,"about_ca_topic_score_gemma":0.00006286753,"teacher_disagreement_score":0.16261344,"about_ca_system_score_codex":0.000009467633,"about_ca_system_score_gemma":0.00009516617,"threshold_uncertainty_score":0.5983721},"labels":[],"label_agreement":null},{"id":"W1967188199","doi":"10.2202/1557-4679.1198","title":"Comparing Approaches to Causal Inference for Longitudinal Data: Inverse Probability Weighting versus Propensity Scores","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":45,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; McGill University","keywords":"Propensity score matching; Causal inference; Inverse probability weighting; Estimator; Covariate; Statistics; Inverse probability; Weighting; Observational study; Average treatment effect; Econometrics; Marginal structural model; Mathematics; Inference; Mean squared error; Global Positioning System; Confidence interval; Computer science; Bayesian probability; Medicine; Posterior probability; Artificial intelligence","score_opus":0.6633736756072357,"score_gpt":0.4530269615981859,"score_spread":0.2103467140090498,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1967188199","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.5040024,0.000004070826,0.49307737,0.0010519302,0.0011024713,0.00040924374,0.00017324119,0.000034900153,0.00014439765],"genre_scores_gemma":[0.61542636,0.0000032807504,0.3841811,0.000035921377,0.0003069726,0.000006996599,0.000016754813,0.000011987142,0.000010611464],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9983362,0.000052510048,0.0006243219,0.00021922866,0.0005642172,0.00020350065],"domain_scores_gemma":[0.99538153,0.0024857146,0.0006306467,0.00048820858,0.00090921053,0.000104656756],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001683063,0.00016757667,0.0002923752,0.00009915435,0.000115149734,0.00014380373,0.0016059981,0.00005746256,0.000016117781],"category_scores_gemma":[0.010886909,0.00011538083,0.00005137013,0.000089827736,0.00020135306,0.00036048063,0.0005902982,0.00049784826,0.0000037006066],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022798565,0.0004488319,0.027645703,0.00012536018,0.0004959269,0.000042239473,0.0011015659,0.00030447336,0.007980762,0.94009507,0.009358838,0.010121368],"study_design_scores_gemma":[0.0016940641,0.00064421527,0.0075008883,0.00025648385,0.00023591403,0.00014683853,0.00024852436,0.016948953,0.020434583,0.9496724,0.0017424407,0.00047471668],"about_ca_topic_score_codex":0.000035921094,"about_ca_topic_score_gemma":0.0011337369,"teacher_disagreement_score":0.11142399,"about_ca_system_score_codex":0.000108864115,"about_ca_system_score_gemma":0.00020447817,"threshold_uncertainty_score":0.9974448},"labels":[],"label_agreement":null},{"id":"W1971149891","doi":"10.2202/1557-4679.1126","title":"Integration of Genetic Familial Dependence Structure in Latent Class Models","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Genomic variations and chromosomal abnormalities","field":"Biochemistry, Genetics and Molecular Biology","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université Laval; Douglas Mental Health University Institute","funders":"National Institute of Mental Health; Canadian Institutes of Health Research","keywords":"Pedigree chart; Latent class model; Disease; Mendelian inheritance; Genetic heterogeneity; Class (philosophy); Genetics; Replicate; Local independence; Conditional independence; Econometrics; Latent variable model; Biology; Statistics; Computer science; Mathematics; Phenotype; Artificial intelligence; Gene; Medicine","score_opus":0.010834150917422828,"score_gpt":0.24257837076877728,"score_spread":0.23174421985135446,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1971149891","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9610436,0.00019126473,0.037780426,0.00045107372,0.0002831699,0.000044104247,0.000114285634,7.6304985e-7,0.00009132792],"genre_scores_gemma":[0.9920525,0.00023898296,0.007294331,0.0001810041,0.00016821828,2.912785e-7,0.000023151768,0.0000036072436,0.00003791332],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99924976,0.000032053053,0.00036769672,0.000061677936,0.00022017403,0.00006861466],"domain_scores_gemma":[0.9992974,0.000017536471,0.00025532884,0.00008370107,0.00032607384,0.00001994973],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00011241549,0.000066353954,0.00008649076,0.000056155637,0.0000157045,0.000017817965,0.0003174473,0.000047270278,0.000010111868],"category_scores_gemma":[0.0000611296,0.000046995585,0.00004234968,0.000035316614,0.00003494757,0.000007492055,0.000028535645,0.00008842556,3.5426888e-7],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00041794876,0.000121960344,0.002922722,0.000005891085,0.00011956524,0.00003200174,0.000833557,0.045907296,0.91560334,0.017462851,0.0007121409,0.01586074],"study_design_scores_gemma":[0.0032045536,0.0020758242,0.28673542,0.00020521568,0.00012434021,0.001061445,0.0015084221,0.029080529,0.49811065,0.17616825,0.0012214348,0.00050391624],"about_ca_topic_score_codex":0.000027858458,"about_ca_topic_score_gemma":0.00007323387,"teacher_disagreement_score":0.41749266,"about_ca_system_score_codex":0.000027243123,"about_ca_system_score_gemma":0.00009065864,"threshold_uncertainty_score":0.19164243},"labels":[],"label_agreement":null},{"id":"W1972903893","doi":"10.2202/1557-4679.1207","title":"Impact of Outcome Model Misspecification on Regression and Doubly-Robust Inverse Probability Weighting to Estimate Causal Effect","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia; Université du Québec à Montréal","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Inverse probability weighting; Covariate; Estimator; Propensity score matching; Statistics; Average treatment effect; Econometrics; Confounding; Outcome (game theory); Regression analysis; Regression; Weighting; Mathematics; Observational study; Inverse probability; Medicine; Bayesian probability","score_opus":0.14815524905517047,"score_gpt":0.470020886576831,"score_spread":0.32186563752166053,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1972903893","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.83579457,0.000002572744,0.16316856,0.00043188862,0.00019381753,0.00019933381,0.00006518187,0.000018857923,0.00012522002],"genre_scores_gemma":[0.7262388,0.0000044290036,0.273621,0.000026067197,0.00007001304,0.0000026276703,0.0000029607688,0.000013102585,0.000021018208],"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99854755,0.00006380797,0.00064088474,0.00012116475,0.0005063777,0.00012020978],"domain_scores_gemma":[0.99699855,0.001201261,0.00077582983,0.00023184529,0.0006981715,0.00009433368],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011061078,0.00015653012,0.00026272002,0.00014221162,0.0000479064,0.000040584287,0.0003960585,0.00006458637,0.000020438396],"category_scores_gemma":[0.0037243653,0.00008932099,0.00007766524,0.00007176698,0.00011600831,0.00012366305,0.00009000976,0.0004006464,0.000001986826],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0036406666,0.0008935545,0.060802836,0.0002489452,0.0005318205,0.0000929983,0.0027745678,0.02843456,0.45408633,0.404622,0.009375091,0.034496658],"study_design_scores_gemma":[0.0012752444,0.0017619476,0.019232778,0.00057744374,0.0001416958,0.00026923497,0.000035852474,0.07887905,0.07507081,0.8223613,0.00003668353,0.00035794708],"about_ca_topic_score_codex":0.000014674922,"about_ca_topic_score_gemma":0.000025318652,"teacher_disagreement_score":0.41773933,"about_ca_system_score_codex":0.00013445878,"about_ca_system_score_gemma":0.00006355269,"threshold_uncertainty_score":0.44586822},"labels":[],"label_agreement":null},{"id":"W1978765295","doi":"10.2202/1557-4679.1275","title":"Sample Size Requirements for Interval Estimation of the Kappa Statistic for Interobserver Agreement Studies with a Binary Outcome and Multiple Raters","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Reliability and Agreement in Measurement","field":"Decision Sciences","cited_by":127,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"","keywords":"Statistics; Cohen's kappa; Confidence interval; Kappa; Intraclass correlation; Mathematics; Interval estimation; Statistic; Sample size determination; Sample (material); Binary number; Limit (mathematics); Interval (graph theory); Psychometrics; Combinatorics","score_opus":0.2216258336390316,"score_gpt":0.43800712383045454,"score_spread":0.21638129019142294,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1978765295","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.68724144,0.000036467747,0.30452096,0.005011971,0.0020031142,0.00057676557,0.0006028786,0.0000021120886,0.0000042817906],"genre_scores_gemma":[0.8858047,0.000012907027,0.11370487,0.00029963904,0.0000769631,0.00001203056,0.000003935047,0.0000068586305,0.00007806447],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9970289,0.0001053058,0.0011673896,0.00016616363,0.0014008649,0.00013136466],"domain_scores_gemma":[0.9857738,0.0104351,0.0012699215,0.0002525876,0.0022239909,0.000044623786],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0041139657,0.00013358652,0.00028243774,0.00007955485,0.00013568788,0.00013081508,0.0009536272,0.000026629055,0.00003126899],"category_scores_gemma":[0.025751788,0.000061026185,0.0001129646,0.000091023794,0.0003308829,0.00017073109,0.00017405467,0.00014133,9.1031797e-7],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.014822012,0.0020801928,0.54771596,0.00074401,0.0062130727,0.000024004725,0.017244702,0.031118978,0.071794756,0.044940155,0.06721252,0.19608964],"study_design_scores_gemma":[0.015584549,0.0061426708,0.27421665,0.0012391439,0.0010017998,0.0001133647,0.014398312,0.17465077,0.019064365,0.48324764,0.009562142,0.0007785992],"about_ca_topic_score_codex":0.000020039623,"about_ca_topic_score_gemma":0.00015474774,"teacher_disagreement_score":0.4383075,"about_ca_system_score_codex":0.00007405372,"about_ca_system_score_gemma":0.0000755367,"threshold_uncertainty_score":0.9824547},"labels":[],"label_agreement":null},{"id":"W1980199310","doi":"10.2202/1557-4679.1400","title":"Special Issue on Causal Inference in Health Research","year":2012,"lang":"en","type":"editorial","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"","keywords":"Causal inference; Biostatistics; Inference; Library science; Data science; Computer science; Econometrics; Mathematics; Public health; Medicine; Artificial intelligence","score_opus":0.26051950652904116,"score_gpt":0.5478536902653409,"score_spread":0.28733418373629976,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1980199310","genre_codex":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000065926826,0.00016811136,0.011320143,0.0035653017,0.9794168,0.00043633903,0.000953701,0.00003309993,0.004040619],"genre_scores_gemma":[0.0011158028,0.0014629719,0.018848402,0.00014376009,0.97729474,0.000008509545,0.000057647485,0.000063131105,0.0010050568],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9931746,0.0006625053,0.0015125666,0.0002028297,0.0039297775,0.00051769643],"domain_scores_gemma":[0.97068435,0.024416545,0.0016408927,0.00036850927,0.0027324825,0.00015721963],"candidate_categories":["metaresearch","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.006832744,0.00030004734,0.0006192147,0.0007415628,0.000090454865,0.00014206028,0.0019301138,0.0003700094,0.00040765907],"category_scores_gemma":[0.039248466,0.00021444878,0.00009549125,0.00023322937,0.00028918093,0.00016523334,0.00029704435,0.0038035284,0.00007577533],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00021690247,0.0001962493,0.000010222497,0.000054915123,0.00007250819,0.00010550283,0.00063725724,0.00001052883,0.0000082515435,0.054233015,0.9326647,0.011789926],"study_design_scores_gemma":[0.00029899736,0.0004599616,0.0000147002365,0.00097607024,0.0000118690405,0.000013891492,0.00013003236,0.000011817873,0.000104104096,0.24252053,0.7553101,0.00014791758],"about_ca_topic_score_codex":0.00014115598,"about_ca_topic_score_gemma":0.00027873012,"teacher_disagreement_score":0.1882875,"about_ca_system_score_codex":0.0014860949,"about_ca_system_score_gemma":0.0014605499,"threshold_uncertainty_score":0.99849474},"labels":[],"label_agreement":null},{"id":"W1981772720","doi":"10.2202/1557-4679.1285","title":"A Tutorial on Methods to Estimating Clinically and Policy-Meaningful Measures of Treatment Effects in Prospective Observational Studies: A Review","year":2011,"lang":"en","type":"review","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":98,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Clinical Evaluative Sciences","funders":"Canadian Institutes of Health Research; Ontario Ministry of Health and Long-Term Care; Institute for Clinical Evaluative Sciences; Heart and Stroke Foundation of Canada","keywords":"Observational study; Covariate; Number needed to treat; Propensity score matching; Medicine; Randomized controlled trial; Average treatment effect; Absolute risk reduction; Relative risk; Regression toward the mean; Treatment effect; Statistics; Confidence interval; Econometrics; Mathematics; Internal medicine","score_opus":0.6067013277402696,"score_gpt":0.62685556926826,"score_spread":0.02015424152799039,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1981772720","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000007792983,0.8636306,0.1331329,0.00027013422,0.00097433967,0.0016438263,0.00015326403,0.000018907613,0.00016822612],"genre_scores_gemma":[0.0000066142243,0.558949,0.44046506,0.00012097997,0.00036756953,0.000052450265,0.0000029169444,0.00002201654,0.000013394512],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9963072,0.00066709204,0.0021342242,0.00021021809,0.00052414817,0.00015710287],"domain_scores_gemma":[0.9805916,0.015091253,0.00274928,0.00023312331,0.0012606645,0.00007408856],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0030242512,0.00035810974,0.0019817736,0.00035346652,0.00003066106,0.00002245351,0.000590878,0.0001064046,0.000005899132],"category_scores_gemma":[0.04849284,0.00020865949,0.00024767275,0.00024793117,0.00013119452,0.00006578911,0.00016310504,0.00041629135,0.0000014607391],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009761901,0.00014454205,0.000007794778,0.0049092053,0.00094691094,0.000028612798,0.00085157447,0.00000356658,0.0000022970623,0.16715106,0.0005716042,0.8252852],"study_design_scores_gemma":[0.0007071448,0.0034910461,0.000026673897,0.16487041,0.0016115977,0.00023359434,0.0000562695,0.000033626093,0.000045991197,0.7509033,0.07763145,0.00038890037],"about_ca_topic_score_codex":0.000020511286,"about_ca_topic_score_gemma":0.000010036999,"teacher_disagreement_score":0.82489634,"about_ca_system_score_codex":0.0010518503,"about_ca_system_score_gemma":0.00075137574,"threshold_uncertainty_score":0.9595221},"labels":[],"label_agreement":null},{"id":"W1982299419","doi":"10.1515/1557-4679.1397","title":"Partial Identification arising from Nondifferential Exposure Misclassification: How Informative are Data on the Unlikely, Maybe, and Likely Exposed?","year":2012,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Health, Environment, Cognitive Aging","field":"Environmental Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"QLT (Canada); University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Classification scheme; Identification (biology); Binary classification; Binary number; Statistics; Econometrics; Computer science; Mathematics; Machine learning","score_opus":0.06139565210296041,"score_gpt":0.28638876363063964,"score_spread":0.22499311152767923,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1982299419","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9209437,0.00009434614,0.060270388,0.016500777,0.0010484633,0.00023053517,0.0006645995,0.000008039737,0.00023916428],"genre_scores_gemma":[0.9967976,0.00017668247,0.001345584,0.0009936233,0.0005192341,0.000003579376,0.00009685759,0.000012387281,0.000054453936],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99805444,0.00022008468,0.0004543123,0.00017547392,0.00089840795,0.00019726266],"domain_scores_gemma":[0.997681,0.0008111551,0.00094451144,0.00039787483,0.00005494906,0.00011048299],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011710433,0.00014147851,0.00012637203,0.000044767916,0.00023879917,0.00022568642,0.00092819665,0.00004957167,0.00030467633],"category_scores_gemma":[0.00060771283,0.00008790946,0.00003061647,0.00006936263,0.00028699264,0.00071992073,0.0003519719,0.00035217335,0.00009779324],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012148028,0.0016810677,0.3738217,0.00004344244,0.0016195629,0.000103291015,0.064257644,0.0005680017,0.07895288,0.0118801,0.10028852,0.365569],"study_design_scores_gemma":[0.0006528418,0.00008546447,0.964844,0.00010156123,0.000118214324,0.000062450534,0.0061095683,0.004709805,0.008400073,0.0018743179,0.012819177,0.00022255967],"about_ca_topic_score_codex":0.000039418,"about_ca_topic_score_gemma":0.0000133885005,"teacher_disagreement_score":0.59102225,"about_ca_system_score_codex":0.00017401366,"about_ca_system_score_gemma":0.00002291496,"threshold_uncertainty_score":0.3584844},"labels":[],"label_agreement":null},{"id":"W1986251233","doi":"10.2202/1557-4679.1305","title":"Modeling Fetal Weight for Gestational Age: A Comparison of a Flexible Multi-level Spline-based Model with Other Approaches","year":2011,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Pregnancy and preeclampsia studies","field":"Medicine","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University Health Centre","funders":"Canadian Institutes of Health Research","keywords":"Gestational age; Pointwise; Mathematics; Fetal weight; Covariance; Fetus; Statistics; Spline (mechanical); Confidence interval; Analysis of covariance; Medicine; Pregnancy; Biology","score_opus":0.3369837490580015,"score_gpt":0.3615435908988037,"score_spread":0.024559841840802232,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1986251233","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.03004615,0.0003783789,0.96803033,0.00056291703,0.00013247559,0.00019718916,0.0003323477,0.000006775918,0.00031341118],"genre_scores_gemma":[0.62032044,0.000009040039,0.37935102,0.00011217459,0.00006426808,0.000007052517,0.000016554244,0.000011933951,0.000107559266],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988446,0.000019808987,0.0004799026,0.000091069596,0.00045141467,0.00011316921],"domain_scores_gemma":[0.9986202,0.0001845926,0.0003638337,0.000099470075,0.00068537786,0.00004651642],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002568466,0.0001197185,0.0002708847,0.00010820327,0.000050823564,0.000010880172,0.00023228499,0.000038338065,0.00001705195],"category_scores_gemma":[0.00010467256,0.00007243142,0.00009476713,0.000047324884,0.00013408154,0.000058320573,0.000025700767,0.00015801415,8.2073524e-7],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.041986287,0.0073275976,0.04608469,0.00068785547,0.009988038,0.00022157957,0.04099604,0.62297297,0.0034431715,0.18198127,0.0028477176,0.041462813],"study_design_scores_gemma":[0.0033125703,0.00042217041,0.0021940114,0.00037095766,0.00028567688,0.00006204166,0.0005439834,0.98616964,0.0028374777,0.0036604863,0.000040176335,0.000100787736],"about_ca_topic_score_codex":0.000022222324,"about_ca_topic_score_gemma":0.000042113028,"teacher_disagreement_score":0.5902743,"about_ca_system_score_codex":0.000038222268,"about_ca_system_score_gemma":0.00022293552,"threshold_uncertainty_score":0.29536676},"labels":[],"label_agreement":null},{"id":"W2011025113","doi":"10.2202/1557-4679.1121","title":"IPCW Estimator for Kendall's Tau under Bivariate Censoring","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":34,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université Laval","funders":"Natural Sciences and Engineering Research Council of Canada; Fonds Québécois de la Recherche sur la Nature et les Technologies","keywords":"Estimator; Mathematics; Bivariate analysis; Statistics; Censoring (clinical trials); Nonparametric statistics; Econometrics; Consistent estimator; Efficient estimator; Minimum-variance unbiased estimator","score_opus":0.13811726294086882,"score_gpt":0.468761519193858,"score_spread":0.3306442562529892,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2011025113","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0016425792,0.000048889786,0.9915486,0.0052089677,0.0009126072,0.00013010274,0.00023273558,0.000014371177,0.0002611691],"genre_scores_gemma":[0.16915742,0.000046364723,0.82951546,0.00064271817,0.00044241478,0.0000017787785,0.000004087322,0.000017292963,0.00017248529],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985612,0.000054279455,0.00060204475,0.00010717488,0.0004825285,0.00019274678],"domain_scores_gemma":[0.99580663,0.0027159776,0.00051481347,0.00013507051,0.00073755905,0.000089952584],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007211591,0.00013649336,0.00024511787,0.00008137242,0.000093435934,0.00007309679,0.0004975588,0.000043328615,0.00003146686],"category_scores_gemma":[0.002979662,0.000090491674,0.00010525128,0.000051820913,0.00006071313,0.00010238192,0.00003419395,0.0001934914,0.0000029035057],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018706903,0.00008695789,0.0000035524185,0.000009501758,0.000116332776,0.00003123967,0.00012714471,0.0005853902,0.0017348662,0.95829064,0.0043977047,0.034429587],"study_design_scores_gemma":[0.00070818886,0.00020888944,0.00021004368,0.00007132033,0.00008294777,0.00013172306,0.00008855533,0.0073100007,0.0011602343,0.98738337,0.0025321196,0.000112605856],"about_ca_topic_score_codex":0.0000021714586,"about_ca_topic_score_gemma":0.000001338937,"teacher_disagreement_score":0.16751485,"about_ca_system_score_codex":0.000090856556,"about_ca_system_score_gemma":0.00007802865,"threshold_uncertainty_score":0.36901432},"labels":[],"label_agreement":null},{"id":"W2021255504","doi":"10.1515/1557-4679.1419","title":"Testing the assumptions for the analysis of survival data arising from a prevalent cohort study with follow-up","year":2012,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Insurance, Mortality, Demography, Risk Management","field":"Social Sciences","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University; Université du Québec à Montréal","funders":"Medical Research Council; Health Canada; Pfizer Canada; University of Ottawa; Pfizer","keywords":"Censoring (clinical trials); Mathematics; Statistics; Truncation (statistics); Counting process; Event (particle physics); Survival analysis; Survival function; Econometrics; Independence (probability theory); Kaplan–Meier estimator","score_opus":0.11640586232519527,"score_gpt":0.3890585567505665,"score_spread":0.2726526944253712,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2021255504","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.90433127,0.0003489982,0.0874624,0.0023744807,0.0026678129,0.00092402915,0.0013959078,0.000010053854,0.0004850497],"genre_scores_gemma":[0.99563545,0.000107115884,0.0035067142,0.000082089406,0.00057308085,0.000008812712,0.000030704454,0.0000085059055,0.000047549],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9973406,0.00035301712,0.0005127873,0.0001158516,0.0014952167,0.0001824873],"domain_scores_gemma":[0.9936226,0.0040984033,0.0008856468,0.00042413775,0.0009255641,0.000043645927],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0060281954,0.00009699606,0.00022369948,0.00012762332,0.00047792564,0.00016578159,0.0018142704,0.00002014824,0.000033358574],"category_scores_gemma":[0.0014603154,0.000047206595,0.00010867115,0.00047045687,0.00030443628,0.00020115775,0.00018467117,0.00016098999,9.872725e-7],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00008855985,0.00021223912,0.9774512,0.000002045299,0.007374125,0.000002272084,0.0051675444,0.0006645909,0.000011268695,0.003908535,0.0008175738,0.00430003],"study_design_scores_gemma":[0.00033296066,0.000051082276,0.98008454,0.00002178099,0.004679027,0.0000010633654,0.009748124,0.002663221,0.0000029833648,0.00069715886,0.0016526816,0.000065358785],"about_ca_topic_score_codex":0.0037683405,"about_ca_topic_score_gemma":0.0071963337,"teacher_disagreement_score":0.09130416,"about_ca_system_score_codex":0.00007607731,"about_ca_system_score_gemma":0.0001035388,"threshold_uncertainty_score":0.56966287},"labels":[],"label_agreement":null},{"id":"W2021583648","doi":"10.2202/1557-4679.1208","title":"Accuracy of Conventional and Marginal Structural Cox Model Estimators: A Simulation Study","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":82,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Natural Gas Technologies Centre; Montreal General Hospital; McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Covariate; Marginal structural model; Proportional hazards model; Estimator; Statistics; Inverse probability weighting; Weighting; Logistic regression; Marginal model; Confounding; Econometrics; Inverse probability; Mathematics; Censoring (clinical trials); Accelerated failure time model; Regression analysis; Computer science; Medicine; Bayesian probability","score_opus":0.0926583015508293,"score_gpt":0.4578373595405563,"score_spread":0.36517905798972705,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2021583648","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7312929,0.0000114315035,0.2679836,0.00020065991,0.00024103562,0.00013877044,0.00008510212,0.000011135016,0.00003537984],"genre_scores_gemma":[0.83878016,0.000006968003,0.16107027,0.000025669378,0.000086098255,0.0000014276062,0.0000033670622,0.000009862305,0.000016151609],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99862456,0.000039744922,0.00058074214,0.00007640163,0.0005956194,0.00008295184],"domain_scores_gemma":[0.9962986,0.0016890004,0.00088683463,0.00012604648,0.0009554895,0.00004404133],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00052072096,0.00010815773,0.00018685819,0.00010906661,0.000045188797,0.000043926604,0.0003747026,0.00003884162,0.00005613436],"category_scores_gemma":[0.0023607847,0.00007471787,0.000044305038,0.000046645982,0.0001646045,0.00020613176,0.00008778385,0.0003043994,5.7503263e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009070318,0.00065424474,0.013304765,0.0000664368,0.0006579107,0.00009074701,0.0027258398,0.01281476,0.031735625,0.92040676,0.0012616826,0.015374201],"study_design_scores_gemma":[0.00072652794,0.00021442979,0.0035517325,0.00004904902,0.00008031103,0.00017353139,0.00024488728,0.2178113,0.0021299247,0.7748844,0.000034385168,0.00009953464],"about_ca_topic_score_codex":0.0000077276545,"about_ca_topic_score_gemma":0.000013670024,"teacher_disagreement_score":0.20499654,"about_ca_system_score_codex":0.000029815252,"about_ca_system_score_gemma":0.00010263148,"threshold_uncertainty_score":0.30469063},"labels":[],"label_agreement":null},{"id":"W2035794422","doi":"10.2202/1557-4679.1096","title":"Properties of the Projected Length of the Curve (PLC) and Area Swept out by the Curve (ASC) Indices for the Receiver Operating Characteristic (SROC) Curve","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Global Cancer Incidence and Screening","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; McMaster University","funders":"","keywords":"Receiver operating characteristic; Area under the curve; Context (archaeology); Statistics; Diagnostic odds ratio; Mathematics; Area under curve; Medicine; Internal medicine","score_opus":0.07648062217362388,"score_gpt":0.32806790220197146,"score_spread":0.25158728002834757,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2035794422","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9684628,0.0025295864,0.001953705,0.024574544,0.00088506506,0.00089403917,0.0004929204,0.000005977266,0.00020132161],"genre_scores_gemma":[0.9971274,0.00037616276,0.00033445316,0.0016339945,0.00027906324,0.000006480188,0.0000066535426,0.0000106901125,0.00022506111],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.99803036,0.00016398341,0.0006605386,0.000112757785,0.00087096146,0.00016139627],"domain_scores_gemma":[0.9968609,0.00066434353,0.001054297,0.00023465102,0.0011504439,0.00003531454],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010259289,0.00015613297,0.00026567816,0.000039884937,0.00024744662,0.00008371026,0.0008848014,0.000049166596,0.000017554079],"category_scores_gemma":[0.0021306705,0.000054884684,0.00013094785,0.00012587887,0.00040089223,0.00011934083,0.00010967667,0.0004036528,5.3906365e-7],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.011669354,0.001408236,0.184804,0.00065231737,0.007941781,0.00008100541,0.10367077,0.000966041,0.32874492,0.013159051,0.1568816,0.19002093],"study_design_scores_gemma":[0.0078028776,0.003725291,0.7519572,0.007893231,0.0027692632,0.0022547273,0.025396448,0.020618835,0.15179001,0.0047853407,0.020222712,0.000784106],"about_ca_topic_score_codex":0.00019535824,"about_ca_topic_score_gemma":0.000100319085,"teacher_disagreement_score":0.56715316,"about_ca_system_score_codex":0.000078106845,"about_ca_system_score_gemma":0.0002712168,"threshold_uncertainty_score":0.25507656},"labels":[],"label_agreement":null},{"id":"W2040731810","doi":"10.1515/ijb-2012-0033","title":"A New Species Abundance Distribution Model Based on Model Combination","year":2013,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Windsor","funders":"","keywords":"Relative abundance distribution; Computer science; Abundance (ecology); Machine learning; Artificial intelligence; Parameterized complexity; Decomposition; Ecology; Data mining; Relative species abundance; Algorithm; Biology","score_opus":0.02357858520210514,"score_gpt":0.25322271284242714,"score_spread":0.229644127640322,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2040731810","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.086937755,0.000010163234,0.88939685,0.013782993,0.00044482385,0.00013292131,0.00046061195,0.000013699727,0.0088201845],"genre_scores_gemma":[0.9935807,0.000037167705,0.003384542,0.001250387,0.00006111073,0.000002464767,0.00010206728,0.000007065689,0.0015744723],"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987102,0.000020194182,0.00028239595,0.0000858708,0.00077852176,0.0001228227],"domain_scores_gemma":[0.9992876,0.00009214883,0.0002768324,0.00010875779,0.00015197654,0.000082685176],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00017576975,0.00009536181,0.00008876359,0.0000272497,0.000075314776,0.0001148311,0.00047304961,0.000031658492,0.008298882],"category_scores_gemma":[0.00015217025,0.00006714732,0.00006003176,0.00007964802,0.00009809849,0.00017859811,0.00006009265,0.0001328942,0.00039715355],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00015341361,0.00029115236,0.0013268759,0.0000032094486,0.000035202684,0.000008860814,0.00027769757,0.22458206,0.0060904343,0.07149881,0.68653363,0.00919863],"study_design_scores_gemma":[0.00065868214,0.00009416873,0.018737555,0.000019285073,0.000014864459,0.000011425883,0.00012388677,0.95638514,0.0013476727,0.018372899,0.004127987,0.00010646383],"about_ca_topic_score_codex":0.000058346774,"about_ca_topic_score_gemma":0.000019497493,"teacher_disagreement_score":0.906643,"about_ca_system_score_codex":0.0006639608,"about_ca_system_score_gemma":0.000044815268,"threshold_uncertainty_score":0.99260765},"labels":[],"label_agreement":null},{"id":"W2041711816","doi":"10.2202/1557-4679.1206","title":"Bayesian Inference for Partially Identified Models","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Bayesian Modeling and Causal Inference","field":"Computer Science","cited_by":52,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Identification (biology); Inference; A priori and a posteriori; Bayesian probability; Frequentist inference; Computer science; Econometrics; Posterior probability; Bayesian inference; Set (abstract data type); Limit (mathematics); Observable; Point estimation; Prior probability; Posterior predictive distribution; Mathematics; Data mining; Statistics; Artificial intelligence; Bayesian linear regression; Physics","score_opus":0.03368752178275909,"score_gpt":0.3163777343436428,"score_spread":0.2826902125608837,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2041711816","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0033945397,0.000020086143,0.98946667,0.0041297125,0.0026869124,0.000060534,0.000030361874,0.000015918427,0.00019527624],"genre_scores_gemma":[0.8097313,0.00002189035,0.18940791,0.00048236924,0.00028273297,0.0000023481346,0.0000021117469,0.000005367744,0.000063970605],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99880815,0.000026270296,0.00042597446,0.00011413165,0.00048201787,0.00014343721],"domain_scores_gemma":[0.997728,0.00049793464,0.00034928508,0.00022458217,0.0011199344,0.00008030125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006777025,0.00009225715,0.00011757373,0.000084160405,0.00007865443,0.0003720683,0.0020791525,0.00004554437,0.000011523624],"category_scores_gemma":[0.00044401723,0.000063988125,0.000075297714,0.000071049704,0.00006907958,0.00037879683,0.00012435064,0.00027688203,0.000005984279],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000037633334,0.00005099306,0.000039069462,0.000003636721,0.0000706701,0.000020384694,0.00045797176,0.0031325729,0.0077576884,0.9421623,0.0027886622,0.04347846],"study_design_scores_gemma":[0.00023164476,0.000058468937,0.000080966325,0.000017845028,0.000011333438,0.000080855185,0.000013207593,0.5115901,0.002792617,0.48427454,0.0007728119,0.00007556491],"about_ca_topic_score_codex":0.000012407323,"about_ca_topic_score_gemma":0.0000420242,"teacher_disagreement_score":0.80633676,"about_ca_system_score_codex":0.000017895458,"about_ca_system_score_gemma":0.00024502302,"threshold_uncertainty_score":0.38636175},"labels":[],"label_agreement":null},{"id":"W2043714498","doi":"10.2202/1557-4679.1177","title":"Interval Estimation of Some Epidemiological Measures of Association","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Windsor; University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Statistics; Confidence interval; Mathematics; Estimator; Relative risk; Interval estimation; Coverage probability; Econometrics","score_opus":0.09281863447476195,"score_gpt":0.4080686945668315,"score_spread":0.3152500600920695,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2043714498","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.10967972,0.00002095247,0.88725126,0.0014304004,0.001196495,0.00006358746,0.00016853515,0.000004845231,0.00018420178],"genre_scores_gemma":[0.5575415,0.000022355021,0.44221535,0.000058124708,0.00014325963,5.569121e-7,0.000001877345,0.0000046521905,0.000012355905],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99788266,0.0002169192,0.0010137204,0.000065484426,0.00072030915,0.00010088018],"domain_scores_gemma":[0.9890428,0.007868879,0.0017598909,0.00011721782,0.0011674595,0.000043735705],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0029747686,0.00009229722,0.00035079714,0.00009149795,0.000022596863,0.000018190447,0.0005037811,0.0000863906,0.000108378474],"category_scores_gemma":[0.03972416,0.000056132474,0.000115003946,0.000062565065,0.00014501333,0.00008497486,0.000056731853,0.00035474842,0.0000021740375],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011891166,0.00013810665,0.0022225778,0.000025946032,0.00020675536,0.0000069383764,0.00020370432,0.000022247237,0.010490629,0.9225164,0.002505242,0.06154254],"study_design_scores_gemma":[0.00028719407,0.0001627433,0.0087131765,0.000082530125,0.00008506473,0.000040916362,0.000048776386,0.0037445321,0.00956391,0.9771106,0.000101418416,0.0000591133],"about_ca_topic_score_codex":0.000018023293,"about_ca_topic_score_gemma":0.000007650974,"teacher_disagreement_score":0.44786176,"about_ca_system_score_codex":0.000050980358,"about_ca_system_score_gemma":0.00008343081,"threshold_uncertainty_score":0.96836466},"labels":[],"label_agreement":null},{"id":"W2044055253","doi":"10.2202/1557-4679.1178","title":"A Comparison of Variable Selection Approaches for Dynamic Treatment Regimes","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"National Institute of Mental Health; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Econometrics; Variable (mathematics); Selection (genetic algorithm); Statistics; Feature selection; Mathematics; Computer science; Artificial intelligence","score_opus":0.14747852943176956,"score_gpt":0.43733046187766667,"score_spread":0.28985193244589713,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2044055253","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04051065,0.000023647568,0.95753086,0.00047590584,0.0007542956,0.00013485101,0.00018604477,0.0000051584157,0.00037860774],"genre_scores_gemma":[0.42364863,0.0000063475177,0.57611537,0.000013124202,0.000098730336,0.0000034433206,0.0000034427553,0.000006911896,0.00010403058],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99899554,0.000047976606,0.0005165549,0.000066767025,0.0002823852,0.00009075127],"domain_scores_gemma":[0.99572533,0.0029267976,0.0006324507,0.000089832734,0.00058955164,0.00003604272],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00054516667,0.00008829464,0.00024264032,0.00006631912,0.00004363764,0.000030443493,0.00028556187,0.000044610035,0.0000576672],"category_scores_gemma":[0.0024810438,0.00005385393,0.00006598301,0.00005619764,0.0000911304,0.00003285248,0.000019482153,0.00014640535,7.619378e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00018278853,0.0003169138,0.0005464419,0.000022454507,0.00023598403,0.00000105658,0.00039389383,0.000047422054,0.006511491,0.9468351,0.0013102286,0.043596253],"study_design_scores_gemma":[0.0006530192,0.0006722007,0.0006867127,0.00004263736,0.00016740448,0.00007383436,0.00026333734,0.0721926,0.010713426,0.91263634,0.0018162548,0.00008224739],"about_ca_topic_score_codex":0.000015253206,"about_ca_topic_score_gemma":0.00002036467,"teacher_disagreement_score":0.38313797,"about_ca_system_score_codex":0.00005295047,"about_ca_system_score_gemma":0.000115198054,"threshold_uncertainty_score":0.29702204},"labels":[],"label_agreement":null},{"id":"W2045128971","doi":"10.2202/1557-4679.1093","title":"A Marginal Mixture Model for Selecting Differentially Expressed Genes across Two Types of Tissue Samples","year":2008,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"York University; Western University","funders":"National Human Genome Research Institute; Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Bayesian hierarchical modeling; Bayesian probability; Marginal model; Data set; Mixture model; Multivariate statistics; Bayesian information criterion; Model selection; Selection (genetic algorithm); Set (abstract data type); Mathematics; Marginal likelihood; Covariance; Biology; Statistics; Computer science; Bayes' theorem; Artificial intelligence; Regression analysis","score_opus":0.03631164427114852,"score_gpt":0.33552823760263933,"score_spread":0.2992165933314908,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2045128971","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.42875898,0.00039543255,0.56988215,0.0003426748,0.0002880437,0.0000658075,0.00024267398,0.0000017069696,0.000022525785],"genre_scores_gemma":[0.9750767,0.00040791143,0.023549255,0.00010525311,0.00039686414,0.000003952927,0.00005207778,0.000011044571,0.0003969537],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9992722,0.000025542338,0.00028376738,0.00009040797,0.00023488114,0.00009316177],"domain_scores_gemma":[0.99876875,0.00005009232,0.00037247568,0.00010046606,0.00067715446,0.00003106261],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00014357726,0.00007847709,0.000106301915,0.000031093517,0.00007185007,0.0000178764,0.0003546652,0.000041564384,0.000007744131],"category_scores_gemma":[0.00015326691,0.000055736164,0.00006213926,0.000029935034,0.00006942502,0.0000046269843,0.00005015738,0.000060101444,3.1909823e-7],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003822648,0.000044504486,0.00035298988,0.000008855184,0.00010496514,0.0000015443557,0.00031938407,0.0026594286,0.9866827,0.00036942895,0.003346979,0.0057269237],"study_design_scores_gemma":[0.001024973,0.00018021204,0.0011323553,0.000032852415,0.00003366433,0.00011593199,0.00014757979,0.008036967,0.97555333,0.0013558127,0.012279905,0.000106383384],"about_ca_topic_score_codex":0.000005947507,"about_ca_topic_score_gemma":0.000014116067,"teacher_disagreement_score":0.5463329,"about_ca_system_score_codex":0.000012978033,"about_ca_system_score_gemma":0.00013088563,"threshold_uncertainty_score":0.22728547},"labels":[],"label_agreement":null},{"id":"W2051129785","doi":"10.2202/1557-4679.1171","title":"Inference in Epidemic Models without Likelihoods","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":194,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Guelph","funders":"Natural Sciences and Engineering Research Council of Canada; Biotechnology and Biological Sciences Research Council; Directorate for Biological Sciences","keywords":"Inference; Markov chain Monte Carlo; Computer science; Monte Carlo method; Econometrics; Data mining; Statistics; Machine learning; Artificial intelligence; Mathematics; Bayesian probability","score_opus":0.0851133174261388,"score_gpt":0.42630224393751154,"score_spread":0.3411889265113727,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2051129785","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009609842,0.00006506276,0.9856867,0.0029914605,0.00042799118,0.00007423157,0.000045579283,0.000008715397,0.001090373],"genre_scores_gemma":[0.5708339,0.00009533219,0.42838806,0.000528735,0.00012853366,6.1365324e-7,8.954589e-7,0.0000057308266,0.00001817748],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9981885,0.00015957338,0.00079636934,0.00010028275,0.00058012165,0.00017516111],"domain_scores_gemma":[0.9950377,0.0037471338,0.0004931925,0.0001519517,0.0004967853,0.000073276504],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012262957,0.00012961164,0.000287851,0.00014817214,0.00002999584,0.00006646725,0.00074068736,0.000053437223,0.000056690013],"category_scores_gemma":[0.006663634,0.00008375521,0.00006200882,0.000120176956,0.00008266373,0.00014212202,0.000044306315,0.00037894538,0.000005624951],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00010414748,0.00011108206,0.000662819,0.000005569557,0.00003681379,0.00008888093,0.00047474774,0.00008775895,0.00035317926,0.90367866,0.0012678594,0.09312851],"study_design_scores_gemma":[0.00041157156,0.0001432799,0.0023784158,0.00016184791,0.000025143996,0.00008718728,0.00006866024,0.016627656,0.00023318002,0.9796422,0.00012240159,0.00009844663],"about_ca_topic_score_codex":0.000016109881,"about_ca_topic_score_gemma":0.000010384336,"teacher_disagreement_score":0.5612241,"about_ca_system_score_codex":0.0000886325,"about_ca_system_score_gemma":0.00013049955,"threshold_uncertainty_score":0.7977474},"labels":[],"label_agreement":null},{"id":"W2054723033","doi":"10.2202/1557-4679.1146","title":"Type I Error Rates, Coverage of Confidence Intervals, and Variance Estimation in Propensity-Score Matched Analyses","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":189,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Clinical Evaluative Sciences","funders":"Canadian Institutes of Health Research; Ontario Ministry of Health and Long-Term Care; Institute for Clinical Evaluative Sciences; Heart and Stroke Foundation of Canada","keywords":"Propensity score matching; Statistics; Confidence interval; Matching (statistics); Sample size determination; Observational study; Hazard ratio; Poisson regression; Type I and type II errors; Mathematics; Selection bias; Poisson distribution; Econometrics; Medicine; Population","score_opus":0.23169680834955775,"score_gpt":0.46242425669277515,"score_spread":0.2307274483432174,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2054723033","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.39586854,0.00013840871,0.60246295,0.0010564016,0.00015187873,0.00014857114,0.00003071132,0.0000124475255,0.00013011583],"genre_scores_gemma":[0.91051656,0.00023231034,0.08901347,0.00017271382,0.000024881945,5.794464e-7,0.0000037371874,0.0000063215,0.000029424473],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99878424,0.000078459205,0.0006175352,0.000079833655,0.0003544423,0.0000854835],"domain_scores_gemma":[0.9977137,0.000642649,0.000756833,0.00012047074,0.0007367491,0.000029611369],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006001221,0.00010323472,0.00027183382,0.00012997858,0.000017860486,0.00004099621,0.00034467236,0.000042134532,0.000022326205],"category_scores_gemma":[0.0022444031,0.0000717439,0.00002881018,0.00012729003,0.00009224318,0.00020690158,0.00004662605,0.00021406425,0.0000010052395],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0017859385,0.00073317846,0.0064249383,0.00017118147,0.00047333742,0.0005352585,0.003938177,0.0026315935,0.13741635,0.78732747,0.003423775,0.055138815],"study_design_scores_gemma":[0.00043275696,0.0004054245,0.017070042,0.00056692294,0.000055671186,0.00021214307,0.00011949825,0.005402871,0.031890642,0.9436967,0.000029388088,0.000117969874],"about_ca_topic_score_codex":0.00002360756,"about_ca_topic_score_gemma":0.000021472366,"teacher_disagreement_score":0.514648,"about_ca_system_score_codex":0.00006783221,"about_ca_system_score_gemma":0.00008127855,"threshold_uncertainty_score":0.29256314},"labels":[],"label_agreement":null},{"id":"W2054829060","doi":"10.2202/1557-4679.1205","title":"Cutting Feedback in Bayesian Regression Adjustment for the Propensity Score","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":57,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Propensity score matching; Bayesian probability; Outcome (game theory); Bayesian linear regression; Statistics; Regression; Econometrics; Bayes' theorem; Confounding; Regression analysis; Posterior probability; Mathematics; Bayesian inference; Computer science","score_opus":0.1513029616466612,"score_gpt":0.41370864719192163,"score_spread":0.26240568554526045,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2054829060","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1672452,0.00009933176,0.8208956,0.008388559,0.002210033,0.000737385,0.00008070514,0.000031640455,0.00031153875],"genre_scores_gemma":[0.8278555,0.00007992966,0.17120446,0.00027957,0.00045434228,0.000010300489,0.0000028458314,0.000017649054,0.00009536627],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9988306,0.00003977483,0.00048641613,0.000078976336,0.00043207366,0.00013216285],"domain_scores_gemma":[0.9965772,0.001960562,0.0006140758,0.00018093419,0.0006332103,0.00003402566],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010745877,0.00010805203,0.00015923125,0.00007212295,0.000074929005,0.000052531726,0.0007145498,0.00004812908,0.000034693556],"category_scores_gemma":[0.0027614892,0.000052665808,0.00006113832,0.000060878607,0.00011123024,0.00010802465,0.000104135965,0.0004655292,0.0000014228756],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007588313,0.00038592474,0.0040741717,0.000066616274,0.0002644867,0.000105247746,0.0018731931,0.00009769273,0.03486011,0.6685491,0.028988417,0.2599762],"study_design_scores_gemma":[0.00077651,0.0001817338,0.0073016984,0.00041612645,0.0000676352,0.00030499217,0.00036644895,0.003634653,0.02593419,0.95810664,0.0027497443,0.00015963515],"about_ca_topic_score_codex":0.000017136486,"about_ca_topic_score_gemma":0.00014779667,"teacher_disagreement_score":0.6606103,"about_ca_system_score_codex":0.000080481164,"about_ca_system_score_gemma":0.00008899316,"threshold_uncertainty_score":0.330596},"labels":[],"label_agreement":null},{"id":"W2055588546","doi":"10.2202/1557-4679.1240","title":"Special Issue on Causal Inference","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"","keywords":"Causal inference; Biostatistics; Inference; Computer science; Library science; Econometrics; Data science; Management science; Mathematics; Engineering; Medicine; Artificial intelligence; Public health","score_opus":0.07737566703058796,"score_gpt":0.4291898214705514,"score_spread":0.35181415443996344,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2055588546","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.29682028,0.000027414822,0.5139879,0.026088787,0.046649117,0.0008462745,0.00076630677,0.00030037505,0.11451357],"genre_scores_gemma":[0.76940686,0.00008263264,0.19819285,0.0012221795,0.03003765,0.0000037991913,0.000007008761,0.000042337448,0.00100471],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985591,0.000037740523,0.00047663765,0.00008231892,0.00071097596,0.0001332021],"domain_scores_gemma":[0.99664414,0.001797336,0.00052821724,0.00019952079,0.0007660053,0.00006476112],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00051862164,0.00012766493,0.00016970215,0.000118702614,0.000050774088,0.00007687051,0.00086027675,0.00006357709,0.0009110947],"category_scores_gemma":[0.0042245104,0.0000830088,0.00005962692,0.000063852545,0.0001579962,0.00011684373,0.000083743456,0.00067474344,0.00005830829],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013356398,0.0001354013,0.00016175838,0.000004642292,0.00008600962,0.00013499559,0.00037281372,0.000009952352,0.006069054,0.8101667,0.14187203,0.040853072],"study_design_scores_gemma":[0.0003109713,0.00027810782,0.00044559131,0.00006567483,0.00003125392,0.00023414608,0.00009601616,0.00009661476,0.019890243,0.856764,0.12164264,0.00014469452],"about_ca_topic_score_codex":0.0000067204505,"about_ca_topic_score_gemma":0.00004880181,"teacher_disagreement_score":0.47258654,"about_ca_system_score_codex":0.000058928308,"about_ca_system_score_gemma":0.00010696412,"threshold_uncertainty_score":0.9975848},"labels":[],"label_agreement":null},{"id":"W2059020892","doi":"10.2202/1557-4679.1179","title":"A Comparison of the Statistical Power of Different Methods for the Analysis of Repeated Cross-Sectional Cluster Randomization Trials with Binary Outcomes","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":13,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Clinical Evaluative Sciences","funders":"Canadian Institutes of Health Research; Ontario Ministry of Health and Long-Term Care; Institute for Clinical Evaluative Sciences; Heart and Stroke Foundation of Canada","keywords":"Statistics; Sample size determination; Cluster randomised controlled trial; Statistical power; Generalized estimating equation; Baseline (sea); Cluster (spacecraft); Mathematics; Random effects model; Marginal model; Randomization; Homogeneity (statistics); Econometrics; Regression analysis; Medicine; Randomized controlled trial; Computer science; Meta-analysis; Internal medicine","score_opus":0.10719052713115843,"score_gpt":0.5200359822670371,"score_spread":0.41284545513587867,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2059020892","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1154236,0.000018718985,0.8818747,0.00045206049,0.0008018474,0.0003489391,0.0010567459,0.0000025287827,0.000020854566],"genre_scores_gemma":[0.5860484,0.000005589807,0.41383952,0.000030640287,0.00003695509,0.0000055754704,0.00000892392,0.000008911249,0.000015478616],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.99588835,0.0008961806,0.0021411953,0.00011811685,0.00083723024,0.00011892405],"domain_scores_gemma":[0.94169486,0.05296041,0.003042792,0.0002916772,0.0019694204,0.00004085335],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0053257197,0.00016116975,0.0010677549,0.0001885597,0.000084648564,0.000049534894,0.0007287997,0.00007779481,0.0002249222],"category_scores_gemma":[0.02768714,0.00006304818,0.00040115212,0.000265371,0.0005819706,0.000042786804,0.00008872937,0.0003146813,1.03821634e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.01024974,0.0009973554,0.23253354,0.0001112567,0.012245567,0.0000023732232,0.001452855,0.0008457049,0.013096327,0.71365887,0.0007650676,0.014041328],"study_design_scores_gemma":[0.005033601,0.00047845743,0.63943267,0.00008677755,0.004416538,0.000025289322,0.00025771218,0.056847904,0.012144684,0.28107318,0.00005188816,0.00015132036],"about_ca_topic_score_codex":0.000019287285,"about_ca_topic_score_gemma":0.000030491046,"teacher_disagreement_score":0.4706248,"about_ca_system_score_codex":0.00002978908,"about_ca_system_score_gemma":0.0001225533,"threshold_uncertainty_score":0.9805031},"labels":[],"label_agreement":null},{"id":"W2074774273","doi":"10.2202/1557-4679.1224","title":"A First Passage Time Model for Long-Term Survivors with Competing Risks","year":2011,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Expectation–maximization algorithm; Inverse Gaussian distribution; Term (time); Gompertz function; Econometrics; Computer science; Maximization; Accelerated failure time model; Statistics; Maximum likelihood; Distribution (mathematics); Mathematics; Covariate; Mathematical optimization","score_opus":0.556491896583013,"score_gpt":0.5140196862311603,"score_spread":0.04247221035185267,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2074774273","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020822993,0.000014925801,0.9766361,0.00061631156,0.0006129757,0.00027713185,0.0004815625,0.000018080624,0.00051992637],"genre_scores_gemma":[0.19791889,0.00002883633,0.8013006,0.00017940543,0.00034312005,0.0000070410256,0.000002785824,0.000037364123,0.00018194178],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99761444,0.00020612818,0.0010878982,0.00014944679,0.0007251673,0.00021691533],"domain_scores_gemma":[0.9687907,0.028615925,0.0012077439,0.00021999022,0.0010559714,0.00010967476],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0032172494,0.00018189798,0.0004418859,0.000086496744,0.000098582284,0.00006652562,0.0009315323,0.00007489442,0.00020314443],"category_scores_gemma":[0.022071183,0.0001099602,0.00014563052,0.000069407426,0.000232967,0.00007370796,0.00010342538,0.00030274235,0.000012389577],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.012303281,0.0026544181,0.02399039,0.00053091213,0.005678669,0.0010097368,0.008952928,0.0016381145,0.0004826435,0.88659745,0.015192767,0.040968668],"study_design_scores_gemma":[0.0017856704,0.0004683873,0.0030977966,0.00035763456,0.00037100646,0.00016570599,0.00007751989,0.056156117,0.00048697286,0.9367658,0.000044926634,0.00022248398],"about_ca_topic_score_codex":0.0000139096155,"about_ca_topic_score_gemma":0.000053655684,"teacher_disagreement_score":0.1770959,"about_ca_system_score_codex":0.00007317375,"about_ca_system_score_gemma":0.000106629675,"threshold_uncertainty_score":0.9861663},"labels":[],"label_agreement":null},{"id":"W2075362134","doi":"10.2202/1557-4679.1209","title":"Balancing and Elimination of Nuisance Variables","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Bayesian Modeling and Causal Inference","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Covariate; Nuisance parameter; Counterfactual thinking; Econometrics; Mathematics; Statistics; Confounding; Propensity score matching; Regression analysis; Regression; Psychology; Estimator","score_opus":0.00920589390074001,"score_gpt":0.256456430879477,"score_spread":0.247250536978737,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2075362134","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0536562,0.00005075413,0.94305223,0.0021140294,0.000900764,0.000014861577,0.0000082644765,0.000004189493,0.0001987203],"genre_scores_gemma":[0.862053,0.00005544481,0.13764945,0.00013018998,0.00009421973,1.6271984e-7,4.3370596e-7,0.0000018964099,0.000015216641],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9993245,0.000022440732,0.00024708617,0.000053114705,0.0002988372,0.00005400397],"domain_scores_gemma":[0.9987254,0.000304799,0.00029453382,0.00009235157,0.00055484223,0.000028077311],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000589711,0.00004424766,0.00007281257,0.00005675495,0.000027267246,0.000071321956,0.00063461374,0.000022296126,0.00000572773],"category_scores_gemma":[0.0003604966,0.000030378911,0.000017373695,0.000052028816,0.00005989419,0.000149784,0.000074621326,0.00015669614,8.9562104e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000015255216,0.000034439858,0.00060155854,0.000006685697,0.000041547777,0.000013470781,0.0006626223,0.00016784703,0.053492874,0.84818864,0.0006652495,0.09610979],"study_design_scores_gemma":[0.0008210542,0.00026341208,0.022201983,0.00021721015,0.00004431116,0.0010485322,0.00014928453,0.504069,0.052422952,0.4165404,0.0019743484,0.00024750616],"about_ca_topic_score_codex":0.000012430629,"about_ca_topic_score_gemma":0.0000053239423,"teacher_disagreement_score":0.80839676,"about_ca_system_score_codex":0.000008384891,"about_ca_system_score_gemma":0.000069462694,"threshold_uncertainty_score":0.1238816},"labels":[],"label_agreement":null},{"id":"W2075916106","doi":"10.2202/1557-4679.1172","title":"Using Generalized Additive Models to Detect and Estimate Threshold Associations","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Statistical Methods and Models","field":"Mathematics","cited_by":31,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Medical Research Council; Canadian Institutes of Health Research; Medical Research Council Canada; McGill University","keywords":"Mathematics; Covariate; Statistics; Threshold model; Curvature; Linear model; Range (aeronautics); Linear regression; Generalized linear model; Econometrics","score_opus":0.20462958845255994,"score_gpt":0.47037034224618895,"score_spread":0.26574075379362905,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2075916106","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.017098041,0.000061265106,0.9806279,0.0011184483,0.00023991153,0.00010668943,0.0005009036,0.000009244022,0.00023760808],"genre_scores_gemma":[0.20063993,0.000053947817,0.7985146,0.000593882,0.00015089403,8.3631323e-7,0.0000030582287,0.000010472333,0.00003236295],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99873155,0.00007968674,0.00046743755,0.00009834859,0.00047095207,0.00015201319],"domain_scores_gemma":[0.9976119,0.0011600782,0.00038033864,0.000092802584,0.0006455796,0.000109310575],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00066057546,0.00011665231,0.00022972257,0.0000949901,0.00009326255,0.00007555673,0.00026713967,0.000036576195,0.000013234523],"category_scores_gemma":[0.0015606026,0.00008149865,0.000051473322,0.00007146451,0.00004621198,0.0001302442,0.00004989136,0.00016776785,9.618767e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011877397,0.000043442917,0.0000063355665,0.000003160715,0.0001736997,0.00006964732,0.00075177365,0.009200078,0.004021373,0.912849,0.0018625429,0.070900135],"study_design_scores_gemma":[0.0003576045,0.00011179759,0.00013726205,0.000047142603,0.000079303536,0.0000961218,0.000051194776,0.066664554,0.00090239936,0.9313769,0.00008285386,0.000092869224],"about_ca_topic_score_codex":0.00000784102,"about_ca_topic_score_gemma":0.000007225431,"teacher_disagreement_score":0.1835419,"about_ca_system_score_codex":0.00011090915,"about_ca_system_score_gemma":0.00005741783,"threshold_uncertainty_score":0.33234185},"labels":[],"label_agreement":null},{"id":"W2079826663","doi":"10.2202/1557-4679.1210","title":"Model Checking with Residuals for g-estimation of Optimal Dynamic Treatment Regimes","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; Health Canada; UNICEF; Thrasher Research Fund","keywords":"Counterfactual thinking; Residual; Context (archaeology); Estimation; Computer science; Econometrics; Mathematics; Statistics; Algorithm; Economics","score_opus":0.08719899470162384,"score_gpt":0.4254682716849469,"score_spread":0.33826927698332304,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2079826663","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.18119255,0.000011397338,0.8176613,0.00062137825,0.00010741546,0.00016287621,0.00013092038,0.000014183085,0.00009800886],"genre_scores_gemma":[0.5014082,0.00002041079,0.49841568,0.000015945823,0.00003208573,0.000004806383,0.000005633375,0.000011115658,0.00008610036],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99900013,0.000015688287,0.00045352033,0.00006989928,0.00036573995,0.00009501206],"domain_scores_gemma":[0.9973002,0.00089590985,0.0008346021,0.0001490736,0.00078992994,0.00003026804],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00037836577,0.000110634835,0.00019775263,0.00009912742,0.000036916463,0.000029057042,0.0003380859,0.00004237781,0.000008830415],"category_scores_gemma":[0.00077455526,0.000065704546,0.00005686406,0.000039071096,0.00011465397,0.000126392,0.000025319088,0.00013509407,3.3324716e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0011213624,0.00040602539,0.00008144346,0.00006274732,0.00065085874,0.000028743389,0.0022191058,0.03797589,0.05993248,0.85562825,0.0016987654,0.040194303],"study_design_scores_gemma":[0.0007683305,0.0006771574,0.000054777312,0.00015047392,0.00014727507,0.00017915519,0.00017717406,0.2744632,0.077863514,0.6453077,0.00009423715,0.00011699002],"about_ca_topic_score_codex":0.000004940126,"about_ca_topic_score_gemma":0.000025640797,"teacher_disagreement_score":0.32021567,"about_ca_system_score_codex":0.00008898136,"about_ca_system_score_gemma":0.00015067398,"threshold_uncertainty_score":0.26793537},"labels":[],"label_agreement":null},{"id":"W2087289223","doi":"10.1515/ijb-2012-0046","title":"Exact Nonparametric Confidence Bands for the Survivor Function","year":2013,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Confidence interval; Nonparametric statistics; Statistics; Mathematics; Estimator; Applied mathematics","score_opus":0.07697518084313971,"score_gpt":0.373519926900873,"score_spread":0.2965447460577333,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2087289223","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007095067,0.0001187745,0.9871038,0.0031159832,0.00189983,0.00023991706,0.00009787018,0.0000071584764,0.00032157532],"genre_scores_gemma":[0.8056937,0.0001280433,0.19247933,0.00073146576,0.00058654073,0.000018964254,0.0000026427551,0.000016546635,0.00034275564],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99856436,0.00009523699,0.000543333,0.00008172943,0.00057186006,0.00014350339],"domain_scores_gemma":[0.9794598,0.018282728,0.00053031364,0.00016230493,0.0015095597,0.00005525398],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0012605998,0.00010792198,0.00017581726,0.00008265328,0.00010578765,0.00017988718,0.00067589857,0.00003862881,0.0005078898],"category_scores_gemma":[0.010378054,0.000053174692,0.00009133214,0.00012398195,0.0001336444,0.00009150982,0.000049334456,0.00020846925,0.000024177796],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00020202844,0.00009979537,0.0003882789,0.00002361946,0.00039343012,0.000005998248,0.00021293884,0.000032458935,0.0008084503,0.81700337,0.06294445,0.11788518],"study_design_scores_gemma":[0.00052439835,0.00032731242,0.009741029,0.00005503829,0.00015901958,0.00008730474,0.00030628667,0.013881286,0.00063276506,0.96971637,0.0044573345,0.00011188427],"about_ca_topic_score_codex":0.00006270066,"about_ca_topic_score_gemma":0.00000654114,"teacher_disagreement_score":0.79859865,"about_ca_system_score_codex":0.000046520192,"about_ca_system_score_gemma":0.00007463301,"threshold_uncertainty_score":0.99795794},"labels":[],"label_agreement":null},{"id":"W2087755581","doi":"10.1515/ijb-2013-0045","title":"Principal Stratification: A Broader Vision","year":2013,"lang":"en","type":"letter","venue":"The International Journal of Biostatistics","topic":"Early Childhood Education and Development","field":"Social Sciences","cited_by":5,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Jewish General Hospital","funders":"","keywords":"Stratification (seeds); Econometrics; Principal (computer security); Statistics; Mathematics; Computer science","score_opus":0.02699953065857512,"score_gpt":0.34259973495747564,"score_spread":0.3156002042989005,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2087755581","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00047827768,0.00017363105,0.0016229515,0.9842423,0.005726399,0.00018514332,0.000046788547,0.0000127901385,0.007511715],"genre_scores_gemma":[0.0286391,0.0014928229,0.017068805,0.8795655,0.037420347,0.000012117929,0.00025064763,0.00005982304,0.035490844],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9970866,0.00020783172,0.0006311139,0.00012177537,0.0017594568,0.0001932536],"domain_scores_gemma":[0.9965975,0.00052148104,0.0009384602,0.00015970036,0.0017048102,0.00007801769],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00076874025,0.00014054305,0.0001710985,0.00016637426,0.00021698332,0.0005385163,0.0013070725,0.00022805491,0.0013923924],"category_scores_gemma":[0.0005998347,0.0000981291,0.00010043921,0.000102842896,0.00025134315,0.00018037076,0.000054482316,0.00097457715,0.00024614338],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000050680665,0.000023606637,0.00004622664,0.0000037405218,0.00012089306,0.000041874555,0.005106324,0.0000030994167,0.0000060879775,0.0062653283,0.98044103,0.007936708],"study_design_scores_gemma":[0.00012863458,0.000031723328,0.0017014194,0.00007944894,0.00003013271,0.000039238494,0.0008833647,0.0000069686716,0.000013304921,0.0084557915,0.9885085,0.00012143693],"about_ca_topic_score_codex":0.00026422818,"about_ca_topic_score_gemma":0.000059188886,"teacher_disagreement_score":0.10467681,"about_ca_system_score_codex":0.00030095005,"about_ca_system_score_gemma":0.0019344963,"threshold_uncertainty_score":0.9995205},"labels":[],"label_agreement":null},{"id":"W2088893370","doi":"10.2202/1557-4679.1125","title":"Fitting Smooth-in-Time Prognostic Risk Functions via Logistic Regression","year":2009,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Health Systems, Economic Evaluations, Quality of Life","field":"Economics, Econometrics and Finance","cited_by":30,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; Fonds Québécois de la Recherche sur la Nature et les Technologies","keywords":"Logistic regression; Hazard; Proportional hazards model; Statistics; Parametric statistics; Econometrics; Regression; Hazard ratio; Regression analysis; Mathematics; Computer science; Confidence interval","score_opus":0.19436973666883683,"score_gpt":0.41585482334367285,"score_spread":0.22148508667483602,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2088893370","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.5807262,0.0028793004,0.2905673,0.116115004,0.0050823092,0.0007784896,0.0012604558,0.00005113184,0.0025398026],"genre_scores_gemma":[0.9873522,0.00019574606,0.0074811056,0.003764293,0.00077650516,0.000003561615,0.000024526727,0.000015289672,0.00038677442],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.995966,0.00024258986,0.0031238704,0.00018463355,0.00026388286,0.00021898132],"domain_scores_gemma":[0.99327683,0.0023823068,0.0037755729,0.00020337701,0.00026676702,0.000095156],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.009185921,0.00013837723,0.00046697186,0.00037065902,0.00012850775,0.00012282825,0.0006473076,0.000073582734,0.00026152507],"category_scores_gemma":[0.011838783,0.00012012897,0.00010469504,0.00014387746,0.0000698793,0.00023555961,0.0000410411,0.00036071337,0.0006155614],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007129207,0.0017214887,0.58309984,0.00013833187,0.0008661448,0.00027308194,0.006285892,0.013916642,0.00031227045,0.11900049,0.2339682,0.039704725],"study_design_scores_gemma":[0.0026611674,0.0006769255,0.5769267,0.0007465013,0.00007151713,0.0003075027,0.0007679593,0.05578295,0.00003110553,0.33723405,0.024229184,0.00056440965],"about_ca_topic_score_codex":0.00012711616,"about_ca_topic_score_gemma":0.00002258741,"teacher_disagreement_score":0.406626,"about_ca_system_score_codex":0.00044532082,"about_ca_system_score_gemma":0.000118090255,"threshold_uncertainty_score":0.99648494},"labels":[],"label_agreement":null},{"id":"W2090545088","doi":"10.2202/1557-4679.1350","title":"Meta-Analysis of Observational Studies with Unmeasured Confounders","year":2012,"lang":"en","type":"review","venue":"The International Journal of Biostatistics","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Observational study; Confounding; Propensity score matching; Sample size determination; Bayesian probability; Meta-analysis; Statistics; Econometrics; Causal inference; Randomized controlled trial; Medicine; Computer science; Mathematics; Internal medicine","score_opus":0.9753175800643589,"score_gpt":0.6456409520288836,"score_spread":0.32967662803547526,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2090545088","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000005792015,0.9739572,0.02318592,0.00070966786,0.0007101448,0.00033685792,0.0008454362,0.0000010064145,0.00024797727],"genre_scores_gemma":[0.00073738326,0.98279834,0.013433801,0.00028652055,0.0002918236,0.00001569481,0.00008854975,0.000027114927,0.0023207846],"study_design_codex":"meta_analysis","study_design_gemma":"meta_analysis","domain_scores_codex":[0.9632477,0.006675846,0.016062582,0.00038957378,0.01343489,0.00018937766],"domain_scores_gemma":[0.915814,0.028593628,0.03932249,0.0016772801,0.01447009,0.00012255249],"candidate_categories":["metaresearch","metaepi_broad","insufficient_payload"],"consensus_categories":["metaresearch","metaepi_broad"],"category_scores_codex":[0.0842443,0.0005333159,0.017142098,0.00156589,0.00007862791,0.00045321937,0.004955946,0.00011016368,0.0041038073],"category_scores_gemma":[0.02165222,0.00017207838,0.011984522,0.0022641947,0.00033068712,0.00021420355,0.00019223223,0.00039179582,0.00012462844],"study_design_candidate":"meta_analysis","study_design_consensus":"meta_analysis","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000015128742,0.00006438994,0.00008557027,0.00035099324,0.9155582,0.000017539112,0.00033744253,0.00025360924,1.4238034e-7,0.012393594,0.011740413,0.05918294],"study_design_scores_gemma":[0.000055093005,0.000028569226,0.000043880736,0.00014458608,0.5123247,0.000073482275,0.0002536886,0.00006494628,2.058362e-7,0.001546023,0.4853514,0.00011340308],"about_ca_topic_score_codex":0.000006039207,"about_ca_topic_score_gemma":0.0000523041,"teacher_disagreement_score":0.473611,"about_ca_system_score_codex":0.00014948413,"about_ca_system_score_gemma":0.000649449,"threshold_uncertainty_score":0.9979102},"labels":[],"label_agreement":null},{"id":"W2091956653","doi":"10.2202/1557-4679.1106","title":"Missing Confounding Data in Marginal Structural Models: A Comparison of Inverse Probability Weighting and Multiple Imputation","year":2008,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":62,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia; McGill University","funders":"","keywords":"Inverse probability weighting; Missing data; Inverse probability; Imputation (statistics); Marginal structural model; Statistics; Weighting; Confounding; Censoring (clinical trials); Computer science; Econometrics; Mathematics; Inverse distance weighting; Estimator; Data mining; Posterior probability; Bayesian probability; Medicine","score_opus":0.33358293714958803,"score_gpt":0.4591081692735752,"score_spread":0.12552523212398714,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2091956653","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7083291,0.00004674148,0.29112524,0.00019551846,0.000102178965,0.00009173891,0.000062991225,0.000008061272,0.000038414957],"genre_scores_gemma":[0.6903223,0.000025304113,0.30958474,0.0000144628475,0.0000357886,2.6929382e-7,0.000009617639,0.0000058095948,0.000001757586],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985723,0.000080137754,0.0007234146,0.00009982578,0.0004270475,0.00009728432],"domain_scores_gemma":[0.9971243,0.0014295607,0.0008296272,0.00015901061,0.0004239461,0.00003354713],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007023558,0.000093640614,0.00023103667,0.00011478369,0.0000545276,0.000025664842,0.000460314,0.000034910885,0.000005866494],"category_scores_gemma":[0.0023202042,0.000069222755,0.000020572996,0.00006788277,0.00020142112,0.0003893088,0.00016948301,0.0002338491,1.2405815e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0021839354,0.0007595311,0.34091747,0.00061300973,0.00069684855,0.0004340625,0.04037927,0.015653918,0.035603162,0.47703153,0.0022780416,0.08344922],"study_design_scores_gemma":[0.0004613806,0.0000605444,0.0020228315,0.00016633423,0.000020304231,0.00023494598,0.0005158538,0.419043,0.0020047745,0.57538563,0.000011849391,0.000072562085],"about_ca_topic_score_codex":0.000062061496,"about_ca_topic_score_gemma":0.000076250224,"teacher_disagreement_score":0.4033891,"about_ca_system_score_codex":0.000113373164,"about_ca_system_score_gemma":0.00009954568,"threshold_uncertainty_score":0.2822822},"labels":[],"label_agreement":null},{"id":"W2093789600","doi":"10.2202/1557-4679.1195","title":"Estimating Multilevel Logistic Regression Models When the Number of Clusters is Low: A Comparison of Different Statistical Software Procedures","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":192,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute for Clinical Evaluative Sciences","funders":"Canadian Institutes of Health Research; Ontario Ministry of Health and Long-Term Care; Institute for Clinical Evaluative Sciences; Heart and Stroke Foundation of Canada","keywords":"Statistics; Multilevel model; Logistic regression; Computer science; Variance (accounting); Statistical model; Hierarchical database model; Software; Bayesian probability; Mathematics; Data mining; Econometrics","score_opus":0.08904626714055525,"score_gpt":0.43281489976916027,"score_spread":0.343768632628605,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2093789600","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.041925285,0.000016732902,0.95561045,0.0006282754,0.000872044,0.00016240706,0.00070530793,0.000007998092,0.00007152057],"genre_scores_gemma":[0.49149463,0.000005038513,0.5083044,0.00006559656,0.00009846888,0.0000022356671,0.0000042655583,0.000013003777,0.00001236436],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9969822,0.00017750186,0.0013593145,0.0001467665,0.0011382042,0.00019597806],"domain_scores_gemma":[0.98740625,0.009350559,0.0016676215,0.0002763694,0.0012075214,0.00009169439],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0007497697,0.00021701424,0.00053403986,0.000070161535,0.00009271872,0.00006146158,0.0009849741,0.00009186198,0.00030307434],"category_scores_gemma":[0.012266641,0.000109203706,0.000111464135,0.00005331852,0.00057089824,0.00007882264,0.00018367452,0.00061904575,0.0000023805014],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000701734,0.0010216885,0.006128696,0.0006475479,0.00055486406,0.00004113012,0.007462082,0.00032355596,0.0023499776,0.8672666,0.014722598,0.09877952],"study_design_scores_gemma":[0.00045646733,0.000093200244,0.0014705192,0.0005519483,0.00014648585,0.00010580387,0.00027487727,0.18710372,0.001841233,0.807835,0.000008280462,0.00011245574],"about_ca_topic_score_codex":0.000036603353,"about_ca_topic_score_gemma":0.000017342943,"teacher_disagreement_score":0.44956934,"about_ca_system_score_codex":0.0000397644,"about_ca_system_score_gemma":0.00017809412,"threshold_uncertainty_score":0.99605346},"labels":[],"label_agreement":null},{"id":"W2095288212","doi":"10.2202/1557-4679.1246","title":"Fast Function-on-Scalar Regression with Penalized Basis Expansions","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Control Systems and Identification","field":"Engineering","cited_by":135,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"National Science Foundation","keywords":"Pointwise; Estimator; Mathematics; Overfitting; Ordinary least squares; Model selection; Applied mathematics; Basis function; Smoothing; Functional data analysis; Algorithm; Statistics; Mathematical optimization; Computer science; Artificial intelligence","score_opus":0.006292328742969971,"score_gpt":0.22358636951406072,"score_spread":0.21729404077109074,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2095288212","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.6948702,0.0001922117,0.28657365,0.003978543,0.011551057,0.00023847737,0.00017991685,0.00007632488,0.002339677],"genre_scores_gemma":[0.99776137,0.00004334471,0.0013611177,0.0000851394,0.00051939714,0.0000026585165,0.000008824272,0.000015109261,0.00020305575],"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","domain_scores_codex":[0.99910563,0.000017632437,0.00029613287,0.000055047294,0.0004497977,0.000075766264],"domain_scores_gemma":[0.99910426,0.00014872485,0.00017042622,0.00013398583,0.00039223308,0.00005034704],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00022858547,0.000086195156,0.00010678054,0.00010190103,0.0000595126,0.00008994309,0.00027351303,0.000033553315,0.00014009741],"category_scores_gemma":[0.00009220938,0.00004807474,0.000046923757,0.000060793365,0.00003299019,0.00008828098,0.000012491878,0.00026377907,0.000020278818],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0016829879,0.00030789597,0.0024300953,0.00004820104,0.0017479567,0.0001978435,0.0015040279,0.019544467,0.6032471,0.04769866,0.15203927,0.16955148],"study_design_scores_gemma":[0.015587891,0.0019412311,0.16287138,0.0022822546,0.0011368472,0.003494847,0.0032658798,0.28134447,0.07282889,0.01762929,0.43570352,0.0019135136],"about_ca_topic_score_codex":0.0000111488525,"about_ca_topic_score_gemma":0.00005084649,"teacher_disagreement_score":0.5304182,"about_ca_system_score_codex":0.000031289113,"about_ca_system_score_gemma":0.000028905815,"threshold_uncertainty_score":0.1960431},"labels":[],"label_agreement":null},{"id":"W2097433879","doi":"10.2202/1557-4679.1223","title":"A Pseudo-EM Algorithm for Clustering Incomplete Longitudinal Data","year":2010,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Guelph","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Cluster analysis; Cholesky decomposition; Expectation–maximization algorithm; Covariance; Computer science; Focus (optics); Missing data; Data mining; Multivariate statistics; Algorithm; Multivariate normal distribution; Determining the number of clusters in a data set; CURE data clustering algorithm; Mathematics; Correlation clustering; Statistics; Artificial intelligence; Maximum likelihood; Machine learning","score_opus":0.05577961052628969,"score_gpt":0.3465629364035154,"score_spread":0.29078332587722566,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2097433879","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00052023283,0.000033057193,0.9925877,0.0028237386,0.0036493319,0.00007697929,0.00023439672,0.000010393346,0.00006416052],"genre_scores_gemma":[0.02878991,0.00002078977,0.96969277,0.000477237,0.0009508662,0.0000012470605,0.000010885412,0.000008661152,0.000047611902],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9988093,0.000041116637,0.00039141366,0.0001623825,0.0004498362,0.0001459517],"domain_scores_gemma":[0.99796146,0.0005337615,0.00037399502,0.0004791998,0.0005790018,0.00007256059],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013311061,0.000100440455,0.00014558705,0.000091819165,0.00008486264,0.00029996716,0.003872291,0.000037285874,0.000009670836],"category_scores_gemma":[0.0003143762,0.000067509485,0.000058615682,0.000070820686,0.000053543266,0.00034546538,0.0006666946,0.00029764898,0.000003088172],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000020812304,0.00003880471,0.00002608369,0.000003994109,0.000117574535,0.000060185484,0.00023353484,0.0000125510915,0.0018872563,0.06987952,0.006347589,0.9213721],"study_design_scores_gemma":[0.0006111206,0.00010582324,0.0005270713,0.00003033645,0.000031879066,0.0013942326,0.000017225022,0.8710575,0.000545651,0.11416264,0.011382232,0.0001342627],"about_ca_topic_score_codex":0.000009613008,"about_ca_topic_score_gemma":0.000038604852,"teacher_disagreement_score":0.9212378,"about_ca_system_score_codex":0.000021645374,"about_ca_system_score_gemma":0.00012276981,"threshold_uncertainty_score":0.71957445},"labels":[],"label_agreement":null},{"id":"W2109185224","doi":"10.1515/ijb-2013-0026","title":"Latent Conditional Individual-Level Models for Infectious Disease Modeling","year":2013,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Animal Disease Management and Epidemiology","field":"Agricultural and Biological Sciences","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Guelph; Brock University","funders":"","keywords":"Covariate; Latent variable; Latent variable model; Population; Computer science; Econometrics; Statistics; Conditional independence; Mathematics; Medicine","score_opus":0.1101053187091808,"score_gpt":0.2829587948001932,"score_spread":0.1728534760910124,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2109185224","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.83742046,0.0002346492,0.14148538,0.017776856,0.00070502795,0.0003728384,0.0017288518,0.000019823996,0.00025612427],"genre_scores_gemma":[0.9962864,0.000092094946,0.0009880918,0.0018389034,0.0004955754,0.00001238918,0.00021819932,0.0000010415246,0.000067308094],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9991511,0.000037831138,0.0003159247,0.0000791169,0.00029168781,0.00012435502],"domain_scores_gemma":[0.9987612,0.00040527002,0.00021079622,0.00002203496,0.00050829345,0.00009243728],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00028299805,0.000078501114,0.00010301216,0.000019898027,0.00009555113,0.000086049215,0.0003919413,0.000024375766,0.00027828047],"category_scores_gemma":[0.00022706891,0.000027921888,0.0001063253,0.000037239093,0.000044439395,0.00016618217,0.00007184659,0.000074654396,0.000017844694],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0010772998,0.0007872542,0.015499863,0.00003413936,0.0017187368,0.000066822446,0.00028901515,0.14859714,0.005859723,0.44669318,0.21535316,0.16402368],"study_design_scores_gemma":[0.00032175845,0.00017580292,0.08119029,0.000020365118,0.000089736575,0.000013323886,0.0001080437,0.32643473,0.0000097646425,0.5903354,0.0011824134,0.000118367425],"about_ca_topic_score_codex":0.00005914518,"about_ca_topic_score_gemma":0.000014800767,"teacher_disagreement_score":0.21417074,"about_ca_system_score_codex":0.000026105125,"about_ca_system_score_gemma":0.000015998921,"threshold_uncertainty_score":0.30469757},"labels":[],"label_agreement":null},{"id":"W2136204757","doi":"10.2202/1557-4679.1349","title":"Double-Robust Estimators: Slightly More Bayesian than Meets the Eye?","year":2012,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Canadian Institutes of Health Research","keywords":"Estimator; Confounding; Bayesian probability; Outcome (game theory); Parametric statistics; Econometrics; Statistics; Mathematics; Nonparametric statistics; Parametric model; Mathematical economics","score_opus":0.12095443698415621,"score_gpt":0.4227510600308651,"score_spread":0.3017966230467089,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2136204757","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.035292316,0.0003624504,0.9426986,0.016740875,0.0020861914,0.00032567367,0.00011798587,0.00009560142,0.0022803184],"genre_scores_gemma":[0.8227566,0.000084830324,0.17550068,0.00047703212,0.0008587123,0.000006246394,0.0000052651176,0.000034229495,0.000276406],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99800545,0.000079584075,0.0006440226,0.00008547159,0.00091162696,0.00027384024],"domain_scores_gemma":[0.99695975,0.0011592609,0.0008098921,0.00030086513,0.00065803545,0.00011222109],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010999839,0.00019317973,0.00022639545,0.000106989784,0.00012378403,0.000109374494,0.0012312963,0.0000666814,0.00010609588],"category_scores_gemma":[0.0008349755,0.0001016303,0.000111705645,0.00011183863,0.0002273123,0.00032813282,0.00016173336,0.00041671473,0.00001653778],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00024415358,0.00024808256,0.0027585414,0.000021310441,0.00043505715,0.000068271314,0.0024738298,0.000118358,0.0015590081,0.9562929,0.032103524,0.003676962],"study_design_scores_gemma":[0.0019201253,0.00032082587,0.006195929,0.00058445864,0.00048908696,0.0017808834,0.0021280635,0.0026844512,0.037547342,0.9117492,0.033873744,0.0007259303],"about_ca_topic_score_codex":0.000017973836,"about_ca_topic_score_gemma":0.0000107710175,"teacher_disagreement_score":0.7874643,"about_ca_system_score_codex":0.00015170805,"about_ca_system_score_gemma":0.00008340324,"threshold_uncertainty_score":0.41443634},"labels":[],"label_agreement":null},{"id":"W2185570897","doi":"10.1515/ijb-2015-0026","title":"Effect of Smoothing in Generalized Linear Mixed Models on the Estimation of Covariance Parameters for Longitudinal Data","year":2015,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Mathematics; Statistics; Smoothing; Generalized estimating equation; Covariance; Covariate; Variance function; Generalized linear model; Semiparametric regression; Generalized linear mixed model; Binary data; Poisson distribution; Estimator; Variance (accounting); Binary number","score_opus":0.33207830972491087,"score_gpt":0.4587031214841648,"score_spread":0.12662481175925394,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2185570897","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.15126568,0.000026617754,0.8469621,0.0006872525,0.0003913102,0.00018335058,0.00045043204,0.0000019436527,0.000031314143],"genre_scores_gemma":[0.52087367,0.0000109297125,0.47901103,0.00004137975,0.000040406936,0.000002543361,0.0000095583955,0.000006993969,0.0000035097753],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9981943,0.00033268478,0.0007136475,0.00009382289,0.000571764,0.000093798764],"domain_scores_gemma":[0.9853911,0.012953386,0.00080935395,0.00027098862,0.00053958263,0.00003562344],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0044285534,0.00010010949,0.0002988723,0.00007836316,0.000020215572,0.000022527853,0.00084111554,0.000035268375,0.0000057941343],"category_scores_gemma":[0.020271888,0.00005383988,0.000049460337,0.00008056623,0.000118873024,0.00008641779,0.00009055981,0.00016003627,4.7117476e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0025078098,0.00017589949,0.00015524251,0.00012385468,0.00029433152,0.0000150211,0.0006920745,0.055620078,0.0004584182,0.90469515,0.0034900133,0.03177208],"study_design_scores_gemma":[0.00088744686,0.0004685729,0.000067424684,0.0002041836,0.00006852102,0.000013632482,0.000042601776,0.581502,0.004835946,0.41185287,0.000011935325,0.00004486579],"about_ca_topic_score_codex":0.00004555915,"about_ca_topic_score_gemma":0.0000062393287,"teacher_disagreement_score":0.52588195,"about_ca_system_score_codex":0.00004932994,"about_ca_system_score_gemma":0.00009885583,"threshold_uncertainty_score":0.9879808},"labels":[],"label_agreement":null},{"id":"W2278177624","doi":"10.1515/ijb-2015-0017","title":"Variable Selection for Confounder Control, Flexible Modeling and Collaborative Targeted Minimum Loss-Based Estimation in Causal Inference","year":2015,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":38,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal; Polytechnique Montréal","funders":"National Institute of Allergy and Infectious Diseases","keywords":"Causal inference; Propensity score matching; Estimator; Computer science; Covariate; Inverse probability weighting; Inverse probability; Average treatment effect; Econometrics; Marginal structural model; Inference; Statistics; Nonparametric statistics; Weighting; Selection bias; Machine learning; Artificial intelligence; Mathematics; Posterior probability","score_opus":0.10003523274824203,"score_gpt":0.4231658834685216,"score_spread":0.3231306507202796,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2278177624","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012231027,0.000044141125,0.98636687,0.00063631637,0.00022149157,0.00027981028,0.00011990714,0.000024583538,0.0000758387],"genre_scores_gemma":[0.5997656,0.0000066415273,0.399994,0.00014248895,0.000046661244,0.00001068298,0.000010044875,0.000009859188,0.000014052348],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99874467,0.00007787067,0.00055431534,0.000097977485,0.00039066066,0.00013452333],"domain_scores_gemma":[0.99525934,0.0016889344,0.00048096242,0.00006815406,0.0024407664,0.00006185321],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010052448,0.000124703,0.00022531093,0.0001714735,0.00004363562,0.000090929665,0.00020590064,0.000065983666,0.000008461917],"category_scores_gemma":[0.004449593,0.000096480726,0.000020807733,0.00015452987,0.00008082453,0.0002641035,0.000022261767,0.00018342196,5.810971e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0014468145,0.00016780247,0.00081311184,0.000038729384,0.00017112067,0.00001282841,0.0007538931,0.2061673,0.0033581334,0.7843611,0.001497768,0.0012113728],"study_design_scores_gemma":[0.0011024503,0.00019701413,0.000013684062,0.00006404777,0.000026849744,0.000018136,0.0001349463,0.5023753,0.0014605802,0.4944896,0.00005250771,0.000064872795],"about_ca_topic_score_codex":0.000032360207,"about_ca_topic_score_gemma":0.000035537574,"teacher_disagreement_score":0.58753455,"about_ca_system_score_codex":0.00025678665,"about_ca_system_score_gemma":0.0005617192,"threshold_uncertainty_score":0.53269},"labels":[],"label_agreement":null},{"id":"W2314463908","doi":"10.1515/ijb-2012-0039","title":"The Effect of Error-in-Confounders on the Estimation of the Causal Parameter When Using Marginal Structural Models and Inverse Probability-of-Treatment Weights: A Simulation Study","year":2014,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Marginal structural model; Econometrics; Weighting; Inverse probability weighting; Statistics; Sign (mathematics); Inverse probability; Confounding; Logistic regression; Mathematics; Empirical research; Empirical likelihood; Estimator; Posterior probability; Bayesian probability; Medicine","score_opus":0.14099169658779223,"score_gpt":0.41563104355077823,"score_spread":0.274639346962986,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2314463908","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.90160656,0.000007663629,0.0974379,0.00025171196,0.00011340215,0.00054362894,0.00001912588,0.0000031653972,0.00001684589],"genre_scores_gemma":[0.984678,0.00000468713,0.015262262,0.000015198846,0.000021749465,0.000004606842,8.0800305e-7,0.000008838139,0.0000038999674],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99826086,0.0004522727,0.00062583556,0.000078791585,0.00049433036,0.00008792367],"domain_scores_gemma":[0.9911359,0.0072933654,0.0010109799,0.00022953644,0.00031364517,0.000016592263],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011579524,0.00013189319,0.00023665237,0.0000660138,0.000075191325,0.00002738097,0.0003543592,0.00003255865,0.0000039906154],"category_scores_gemma":[0.0017418825,0.000052543317,0.00006574457,0.000060967053,0.0002993143,0.00011099589,0.00006576607,0.00015824125,7.378049e-8],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0022100294,0.00031511704,0.016729398,0.00012868167,0.0008899917,0.0000052747832,0.010571672,0.60506034,0.001482506,0.34588888,0.000042362684,0.016675735],"study_design_scores_gemma":[0.00043041425,0.0008114385,0.0011013023,0.00011084526,0.000088867826,0.0000090631265,0.0002214129,0.4728561,0.002396326,0.5219357,0.0000012414566,0.00003727943],"about_ca_topic_score_codex":0.0000908678,"about_ca_topic_score_gemma":0.000098710305,"teacher_disagreement_score":0.17604682,"about_ca_system_score_codex":0.00020459986,"about_ca_system_score_gemma":0.000059326576,"threshold_uncertainty_score":0.21426544},"labels":[],"label_agreement":null},{"id":"W2400611004","doi":"10.1515/ijb-2015-0015","title":"Influence Re-weighted G-Estimation","year":2015,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada; McGill University","keywords":"Estimator; Weighting; Estimation; Parametric statistics; Computer science; Robust statistics; Marginal structural model; Average treatment effect; Econometrics; Mathematical optimization; Inverse probability weighting; Outcome (game theory); Mathematics; Parametric model; Statistics; Causal inference; Medicine; Economics","score_opus":0.14528751352761418,"score_gpt":0.43528432890430363,"score_spread":0.2899968153766894,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2400611004","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.18779936,0.000054247677,0.8074397,0.0028496669,0.0005134274,0.00011057459,0.00004034986,0.00005626744,0.0011364244],"genre_scores_gemma":[0.71744144,0.000028479399,0.28198382,0.00030871195,0.00014315407,0.0000016017735,0.0000036357335,0.000011730446,0.00007745501],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9984917,0.000061154635,0.00052059756,0.00006237298,0.00076238386,0.00010178128],"domain_scores_gemma":[0.9966414,0.00077780534,0.0006551127,0.00015365613,0.0016933373,0.000078719015],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00080859667,0.00009773357,0.00014449905,0.0001070136,0.000029446042,0.000060339757,0.0006504969,0.000040303745,0.000023562368],"category_scores_gemma":[0.0042263353,0.000064066146,0.000038505088,0.00009412574,0.00010163408,0.00026421132,0.00007746007,0.00021472588,0.00001681094],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003394343,0.00020639111,0.0005891185,0.00001867397,0.00026472934,0.0003007232,0.0024545745,0.0023250005,0.002696852,0.8793716,0.08583302,0.025599904],"study_design_scores_gemma":[0.0003078086,0.00014169044,0.00017798682,0.000084769694,0.000032135355,0.00025190017,0.00019470569,0.0034003323,0.007648514,0.98571175,0.0019612373,0.00008715169],"about_ca_topic_score_codex":0.0000101047635,"about_ca_topic_score_gemma":0.0000066293537,"teacher_disagreement_score":0.52964205,"about_ca_system_score_codex":0.00017292675,"about_ca_system_score_gemma":0.00015064448,"threshold_uncertainty_score":0.5059624},"labels":[],"label_agreement":null},{"id":"W2407739463","doi":"10.1515/ijb-2015-0016","title":"The Orthogonally Partitioned EM Algorithm: Extending the EM Algorithm for Algorithmic Stability and Bias Correction Due to Imperfect Data","year":2016,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Algorithm; Stability (learning theory); Computer science; Extension (predicate logic); Missing data; Convergence (economics); Reduction (mathematics); Series (stratigraphy); Expectation–maximization algorithm; Software; Sequence (biology); Mathematics; Maximum likelihood; Statistics","score_opus":0.12748056096607074,"score_gpt":0.3903983024897696,"score_spread":0.26291774152369884,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2407739463","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.006855493,0.00011546037,0.9836891,0.004152598,0.0025317813,0.00039225566,0.0022316887,0.000014016766,0.000017606437],"genre_scores_gemma":[0.06284767,0.00048613313,0.9344238,0.00051586,0.001429099,0.000040526367,0.000024699099,0.00004438355,0.00018784506],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9975315,0.00037774292,0.0008444931,0.0002361226,0.0007439615,0.00026618666],"domain_scores_gemma":[0.978102,0.019568862,0.00061476586,0.00045723768,0.001137265,0.00011985595],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.004913269,0.00018876897,0.0002753516,0.000064734464,0.00041950002,0.00030730193,0.0010920613,0.000051487692,0.000061544546],"category_scores_gemma":[0.01400875,0.00007920465,0.00008033807,0.00011561882,0.00021759035,0.00015932173,0.0003189752,0.00023494945,0.0000058336914],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000105863226,0.0000521968,0.000046525063,0.00000629158,0.00018490583,0.000015121221,0.00034192114,9.214387e-7,0.00029660773,0.019184677,0.0078356555,0.9719293],"study_design_scores_gemma":[0.0019851734,0.0010705944,0.0077928076,0.00046696054,0.00042028626,0.0014467592,0.0029019488,0.11964183,0.003432423,0.8473616,0.012977842,0.0005017586],"about_ca_topic_score_codex":0.000014919088,"about_ca_topic_score_gemma":0.00007915114,"teacher_disagreement_score":0.97142756,"about_ca_system_score_codex":0.00012102721,"about_ca_system_score_gemma":0.00015517921,"threshold_uncertainty_score":0.99429667},"labels":[],"label_agreement":null},{"id":"W249493256","doi":"10.1515/ijb-2014-0037","title":"Double Bias: Estimation of Causal Effects from Length-Biased Samples in the Presence of Confounding","year":2015,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University","funders":"","keywords":"Propensity score matching; Confounding; Observational study; Estimator; Statistics; Econometrics; Hazard ratio; Causal inference; Sample size determination; Proportional hazards model; Sampling bias; Survival analysis; Estimation; Information bias; Mathematics; Selection bias; Point estimation; Data set; Pooling; Confidence interval; Computer science; Economics","score_opus":0.24735357801031266,"score_gpt":0.4444989250048111,"score_spread":0.19714534699449843,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W249493256","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.5802997,0.000081544735,0.41836533,0.00042082398,0.00036002827,0.00019778969,0.0001084546,0.000009290727,0.00015706844],"genre_scores_gemma":[0.8878138,0.00003232187,0.112001054,0.00004692243,0.000077166915,0.0000032726953,0.0000099381705,0.00000885194,0.0000066649663],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.998096,0.00017290821,0.0007267277,0.0000673302,0.0008402273,0.00009679366],"domain_scores_gemma":[0.9905207,0.007465536,0.0010943383,0.00018341866,0.000703831,0.000032178978],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014990823,0.00010393613,0.0002533886,0.00013005841,0.000017785462,0.000033231863,0.0007414512,0.000043399617,0.00000959512],"category_scores_gemma":[0.006737386,0.00006248528,0.00004767459,0.00011363165,0.00016527495,0.00017374192,0.0000661576,0.00021194926,0.0000012382183],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00162481,0.00070645806,0.004854632,0.00018700668,0.0005315774,0.00017198185,0.024492852,0.0057913084,0.024798065,0.91251254,0.007957573,0.016371194],"study_design_scores_gemma":[0.0013578088,0.00031314095,0.0013477269,0.0006313584,0.00008360719,0.000056721998,0.0017686695,0.006158302,0.09024844,0.8978259,0.00010449572,0.00010383237],"about_ca_topic_score_codex":0.00031966044,"about_ca_topic_score_gemma":0.00006840954,"teacher_disagreement_score":0.30751416,"about_ca_system_score_codex":0.00009089945,"about_ca_system_score_gemma":0.0001421564,"threshold_uncertainty_score":0.80657667},"labels":[],"label_agreement":null},{"id":"W2557277836","doi":"10.1515/ijb-2015-0034","title":"Effect Estimation in Point-Exposure Studies with Binary Outcomes and High-Dimensional Covariate Data – A Comparison of Targeted Maximum Likelihood Estimation and Inverse Probability of Treatment Weighting","year":2016,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":21,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University Health Centre; Université de Montréal; McGill University; Jewish General Hospital","funders":"Canadian Institutes of Health Research","keywords":"Covariate; Inverse probability weighting; Estimator; Statistics; Marginal structural model; Average treatment effect; Inverse probability; Econometrics; Weighting; Computer science; Causal inference; Mathematics; Medicine; Bayesian probability","score_opus":0.10122401502630411,"score_gpt":0.4121929389408524,"score_spread":0.3109689239145483,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2557277836","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8809781,0.00010860622,0.11766705,0.0007077931,0.00006539503,0.00025831454,0.0002029454,0.000009692081,0.0000020761486],"genre_scores_gemma":[0.7159647,0.000054807206,0.28394124,0.000007869368,0.0000069951207,0.0000031891798,0.000012571003,0.0000069515663,0.0000016878724],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99853796,0.00016099363,0.0007275989,0.00012963008,0.00035445587,0.00008938744],"domain_scores_gemma":[0.99513984,0.0031199483,0.0010418325,0.00019577885,0.00047207033,0.000030560754],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010170378,0.00014686595,0.00044381004,0.00013782032,0.000028210721,0.0000116130705,0.00020283139,0.000037617327,0.0000040484024],"category_scores_gemma":[0.0023473706,0.00007150096,0.000019978132,0.000066246896,0.00023624557,0.00027096536,0.0001480878,0.000084830295,1.7159061e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0071126875,0.0023402679,0.3030027,0.0013027407,0.0041766576,0.00013962612,0.008414017,0.0036484434,0.04894218,0.06783859,0.00064626837,0.5524358],"study_design_scores_gemma":[0.0051680775,0.0063844905,0.026993869,0.0020471693,0.00046835694,0.0001244142,0.00035264835,0.07806754,0.03755126,0.84258676,0.000003845118,0.00025157887],"about_ca_topic_score_codex":0.000042589665,"about_ca_topic_score_gemma":0.00007949206,"teacher_disagreement_score":0.77474815,"about_ca_system_score_codex":0.00015493174,"about_ca_system_score_gemma":0.00008276525,"threshold_uncertainty_score":0.29157245},"labels":[],"label_agreement":null},{"id":"W2904405044","doi":"10.1515/ijb-2017-0002","title":"Parametric Regression Analysis with Covariate Misclassification in Main Study/Validation Study Designs","year":2018,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"National Institute of Environmental Health Sciences; National Cancer Institute; Natural Sciences and Engineering Research Council of Canada; National Institutes of Health","keywords":"Covariate; Inference; Computer science; Statistics; Statistical inference; Observational error; Data mining; Econometrics; Causal inference; Parametric statistics; Nonparametric statistics; Machine learning; Mathematics; Artificial intelligence","score_opus":0.16327130706442697,"score_gpt":0.4423361246666251,"score_spread":0.27906481760219815,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2904405044","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.34564137,0.000004002559,0.6536016,0.0002306283,0.00017566716,0.00024121624,0.000031212654,0.000005363186,0.00006894094],"genre_scores_gemma":[0.7411235,0.000005326479,0.25869057,0.000027257127,0.00011198009,0.0000053667372,0.000003442581,0.00000971952,0.00002283036],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99713564,0.00072830584,0.00082714914,0.00017407654,0.0009903173,0.00014450222],"domain_scores_gemma":[0.9947984,0.0030239944,0.0008877973,0.00025168408,0.0009792111,0.000058902595],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0028409665,0.00014876966,0.00033956635,0.0006513054,0.000083448016,0.00014375709,0.0005710092,0.000037194903,0.000113033144],"category_scores_gemma":[0.0035807379,0.00008372525,0.000051447096,0.0011132479,0.00011661437,0.000101569116,0.000052771436,0.0002307734,0.000005958093],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005356433,0.012058509,0.62494665,0.000040726958,0.008279947,0.0013664764,0.018645432,0.00029110792,0.0019874694,0.22379145,0.0029307862,0.10030502],"study_design_scores_gemma":[0.0035942916,0.0043319184,0.54836655,0.00016735452,0.0022806532,0.00010422956,0.0077087926,0.010359249,0.0013339471,0.4213593,0.00003383227,0.00035990405],"about_ca_topic_score_codex":0.00006521816,"about_ca_topic_score_gemma":0.00022761268,"teacher_disagreement_score":0.39548212,"about_ca_system_score_codex":0.00015285301,"about_ca_system_score_gemma":0.00011098333,"threshold_uncertainty_score":0.42867362},"labels":[],"label_agreement":null},{"id":"W2924255108","doi":"10.1515/ijb-2018-0090","title":"A Joint Poisson State-Space Modelling Approach to Analysis of Binomial Series with Random Cluster Sizes","year":2019,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of New Brunswick","funders":"","keywords":"Mathematics; Negative binomial distribution; Poisson distribution; Overdispersion; Statistics; Randomness; Series (stratigraphy); Binomial (polynomial); Binomial distribution; State space; Random effects model","score_opus":0.04154293692951882,"score_gpt":0.313282682610338,"score_spread":0.27173974568081916,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2924255108","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.103351235,0.000013765091,0.8949189,0.00060804636,0.00025555247,0.00017153117,0.00025016957,0.000005202387,0.00042558755],"genre_scores_gemma":[0.37642804,0.000014725435,0.6232645,0.00008998301,0.0000684817,0.0000017318123,0.0000041868166,0.000012485726,0.00011589832],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99806166,0.0001496755,0.0007220904,0.00013459985,0.0007761656,0.00015578336],"domain_scores_gemma":[0.9958196,0.0018637768,0.0007096303,0.00019354581,0.0013279924,0.000085417414],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00096505915,0.00015347797,0.0005748361,0.0003210004,0.000031512882,0.0000834321,0.00043628202,0.000035929275,0.00006784764],"category_scores_gemma":[0.0011309303,0.00008853281,0.00013553559,0.0003162021,0.00009241988,0.00009820442,0.000075983895,0.00019412382,0.000003836653],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.012376722,0.00072728755,0.0021408205,0.00022436329,0.010055127,0.00005510045,0.009628584,0.15149756,0.0036444261,0.7985489,0.003796221,0.0073049385],"study_design_scores_gemma":[0.005534595,0.0014728406,0.0025531407,0.0004839582,0.0031170445,0.00015932498,0.0017445098,0.3623632,0.008676775,0.6127387,0.00053772307,0.00061817135],"about_ca_topic_score_codex":0.000052733212,"about_ca_topic_score_gemma":0.000015241373,"teacher_disagreement_score":0.2730768,"about_ca_system_score_codex":0.00006321331,"about_ca_system_score_gemma":0.0000991696,"threshold_uncertainty_score":0.36102632},"labels":[],"label_agreement":null},{"id":"W2996445992","doi":"10.1515/ijb-2017-0092","title":"Incorporating Contact Network Uncertainty in Individual Level Models of Infectious Disease using Approximate Bayesian Computation","year":2019,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Markov Chains and Monte Carlo Methods","field":"Mathematics","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Markov chain Monte Carlo; Approximate Bayesian computation; Computer science; Computation; Bayesian probability; Particle filter; Population; Bayesian network; Monte Carlo method; Variable-order Bayesian network; Markov chain; Context (archaeology); Bayesian inference; Statistics; Machine learning; Artificial intelligence; Algorithm; Mathematics; Inference; Medicine","score_opus":0.10958220674951051,"score_gpt":0.36164546272418807,"score_spread":0.25206325597467755,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2996445992","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.39427343,0.00004063445,0.60463405,0.000086525324,0.00054378394,0.00015135542,0.00012382296,0.0000043747345,0.00014200299],"genre_scores_gemma":[0.8837333,0.000012267389,0.11595831,0.000089411034,0.00016541626,9.596638e-7,0.000013965015,0.000016303533,0.000010073632],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.998083,0.00023591824,0.00082757004,0.00009864919,0.00061257585,0.00014231111],"domain_scores_gemma":[0.99683243,0.0011676729,0.0012450893,0.00011430053,0.0005753969,0.000065121385],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0017773148,0.00013150154,0.00030839772,0.0001597668,0.00003481811,0.000051817402,0.00034461662,0.00004539261,0.000006753087],"category_scores_gemma":[0.00054713216,0.00009835112,0.00009042909,0.00014021918,0.000048134447,0.00014624117,0.00010184854,0.00023874224,3.3196315e-8],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00059191824,0.00023561687,0.018048322,0.00015022827,0.00039146666,0.000094302515,0.0017190388,0.7670748,0.0010504939,0.2025697,0.00040590597,0.007668185],"study_design_scores_gemma":[0.0009905769,0.00008491267,0.00082865625,0.00034053373,0.00006834938,0.0000474453,0.00031141925,0.7149246,0.00006647181,0.28222615,0.0000048449697,0.000106089006],"about_ca_topic_score_codex":0.000079712896,"about_ca_topic_score_gemma":0.000054336724,"teacher_disagreement_score":0.48945987,"about_ca_system_score_codex":0.00014849105,"about_ca_system_score_gemma":0.00021761263,"threshold_uncertainty_score":0.40106422},"labels":[],"label_agreement":null},{"id":"W3008593490","doi":"10.1515/ijb-2019-0038","title":"A Bayesian Framework for Robust Quantitative Trait Locus Mapping and Outlier Detection","year":2020,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Genetic and phenotypic traits in livestock","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Outlier; Bayesian probability; Lasso (programming language); Quantitative trait locus; Anomaly detection; Artificial intelligence; Prior probability; Pattern recognition (psychology); Regression; Computer science; Mean-shift; Mathematics; Statistics; Biology; Genetics","score_opus":0.032881268497582654,"score_gpt":0.28072459969188274,"score_spread":0.24784333119430008,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3008593490","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.042897154,0.0002916803,0.9536501,0.0025500378,0.0003693063,0.000087694934,0.00010913063,0.0000020106734,0.000042879004],"genre_scores_gemma":[0.64562225,0.00003301553,0.35321048,0.00066632486,0.00043709512,0.0000016011434,0.0000067912133,0.000008311989,0.000014156743],"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99942887,0.000027812755,0.00022470894,0.00009495501,0.00014516707,0.00007846323],"domain_scores_gemma":[0.9993234,0.00012257212,0.00019908164,0.000049321137,0.0002463466,0.000059262715],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00013031252,0.00007683557,0.000087816115,0.000022432698,0.000047741745,0.00003576252,0.00021546781,0.00005413272,0.000008861953],"category_scores_gemma":[0.0005608192,0.00005769078,0.000048697973,0.00002987913,0.000075639175,0.0000034131676,0.00003780078,0.00010649379,6.890731e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0064402353,0.0003233754,0.0013931962,0.00018010558,0.0034114046,0.000028057491,0.013644228,0.028555669,0.32119897,0.31607732,0.018692933,0.29005453],"study_design_scores_gemma":[0.009194749,0.015450245,0.0513893,0.0004898626,0.000736511,0.0010797646,0.0152844,0.035862546,0.13676128,0.6291842,0.102839544,0.0017276162],"about_ca_topic_score_codex":0.000002058826,"about_ca_topic_score_gemma":0.0000037638176,"teacher_disagreement_score":0.6027251,"about_ca_system_score_codex":0.000007671852,"about_ca_system_score_gemma":0.000043440858,"threshold_uncertainty_score":0.23525618},"labels":[],"label_agreement":null},{"id":"W3048275956","doi":"10.1515/ijb-2018-0109","title":"Shrinkage estimation applied to a semi-nonparametric regression model","year":2020,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Dam Engineering and Safety","field":"Engineering","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Brock University; Western University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Shrinkage; Estimation; Nonparametric statistics; Econometrics; Statistics; Nonparametric regression; Regression; Regression analysis; Mathematics; Economics","score_opus":0.014250118074654506,"score_gpt":0.24418294687060674,"score_spread":0.22993282879595223,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3048275956","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02194006,0.000055756766,0.9749674,0.0015796124,0.0004497689,0.00005478113,0.00005539776,0.000048896563,0.00084829127],"genre_scores_gemma":[0.93550795,0.00005053105,0.06379074,0.00042908438,0.00017967973,0.000001237521,0.0000067588435,0.000017195223,0.000016815466],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991612,0.0000061316564,0.00029820265,0.000054533106,0.00038916513,0.000090806556],"domain_scores_gemma":[0.9995153,0.00011284212,0.00007370313,0.000071159644,0.00012345701,0.00010351575],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000143052,0.00009301862,0.00011929254,0.00010902403,0.000020334744,0.00004700495,0.00038440773,0.00002863649,0.00001196748],"category_scores_gemma":[0.00022622642,0.00006618096,0.00003718663,0.0001693187,0.000010980433,0.000051986186,0.00003908646,0.0001893728,0.000031402164],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000036481208,0.0000067491064,0.000004574818,0.000009883416,0.000049718277,0.000013229847,0.00043518122,0.9625428,0.0047251163,0.0017418532,0.01056688,0.019867547],"study_design_scores_gemma":[0.00021262102,0.000031228337,0.00016107991,0.000036099827,0.000016358792,0.00002401875,0.000032503314,0.9943076,0.0028442235,0.00096066605,0.0012933516,0.00008029053],"about_ca_topic_score_codex":7.0918287e-7,"about_ca_topic_score_gemma":2.1325074e-7,"teacher_disagreement_score":0.9135679,"about_ca_system_score_codex":0.00006832973,"about_ca_system_score_gemma":0.000022742557,"threshold_uncertainty_score":0.26987815},"labels":[],"label_agreement":null},{"id":"W3093766204","doi":"10.1515/ijb-2020-0042","title":"Parametric models for combined failure time data from an incident cohort study and a prevalent cohort study with follow-up","year":2020,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Cohort; Statistics; Parametric statistics; Statistical inference; Inference; Cohort study; Econometrics; Survival analysis; Duration (music); Proportional hazards model; Medicine; Computer science; Mathematics; Artificial intelligence","score_opus":0.12166238620044295,"score_gpt":0.390560009340417,"score_spread":0.26889762313997406,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3093766204","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.57344013,0.000017796,0.423222,0.00064413244,0.00015103786,0.0011997878,0.0012982325,0.000010868443,0.000016041457],"genre_scores_gemma":[0.8238212,0.000015304397,0.17569038,0.00019297507,0.00015619598,0.000019858311,0.000057141446,0.00002477671,0.000022173303],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9973305,0.00029721417,0.0007590131,0.00031817655,0.0011484413,0.00014667315],"domain_scores_gemma":[0.99415535,0.003720774,0.00065190904,0.00041302835,0.00088050414,0.000178412],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016320567,0.00019941649,0.00050088775,0.00007979307,0.00008230474,0.00020164583,0.0011937338,0.000034899134,0.000077817305],"category_scores_gemma":[0.004303226,0.00012391292,0.00003340357,0.00013069897,0.00007979932,0.00020529733,0.0003199544,0.00024472596,0.0000031639124],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.009975201,0.007207665,0.88848996,0.00010642374,0.011415813,0.0005636089,0.009936891,0.0004035201,0.00020664884,0.03481602,0.025072372,0.011805892],"study_design_scores_gemma":[0.009494572,0.012586733,0.2505035,0.00013375169,0.0033358075,0.00007379684,0.004852626,0.45360932,0.00005095766,0.26472613,0.000101658436,0.0005311349],"about_ca_topic_score_codex":0.000097206044,"about_ca_topic_score_gemma":0.00007976822,"teacher_disagreement_score":0.6379864,"about_ca_system_score_codex":0.000063216845,"about_ca_system_score_gemma":0.000112027934,"threshold_uncertainty_score":0.5151674},"labels":[],"label_agreement":null},{"id":"W3107334064","doi":"10.1515/ijb-2020-0073","title":"Doubly robust adaptive LASSO for effect modifier discovery","year":2022,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University; Université de Montréal","funders":"Canadian Institutes of Health Research","keywords":"Counterfactual thinking; Lasso (programming language); Propensity score matching; Outcome (game theory); Observational study; Covariate; Marginal structural model; Set (abstract data type); Conditional expectation","score_opus":0.13319771240494277,"score_gpt":0.3993879005224699,"score_spread":0.26619018811752715,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3107334064","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02177048,0.000051889056,0.97455585,0.0012186297,0.0010136158,0.00030430182,0.00070450065,0.000026864374,0.0003538509],"genre_scores_gemma":[0.8545878,0.000019525894,0.14382292,0.00035554465,0.00036604912,0.000049529423,0.000017606351,0.0000338513,0.000747205],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9984675,0.000116592964,0.00044619027,0.00009463931,0.000730436,0.00014463472],"domain_scores_gemma":[0.99550855,0.003198523,0.0006496213,0.00014315463,0.0004643261,0.00003584484],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010792502,0.0001260066,0.00021473592,0.00011127144,0.00013620203,0.00007160407,0.00077759934,0.000025002859,0.000060564573],"category_scores_gemma":[0.0011917577,0.00008586997,0.00012842425,0.00007201898,0.00007657733,0.0001947049,0.00020114117,0.0003354802,0.0000013991316],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0020620837,0.00023763807,0.00011210992,0.000024104298,0.000605076,0.00012566669,0.00067867234,0.006260634,0.0032611047,0.8992999,0.078996636,0.008336353],"study_design_scores_gemma":[0.0011415072,0.0015186287,0.000062314975,0.000053425927,0.00013563187,0.00033361663,0.00037353337,0.0046208673,0.009064841,0.97609055,0.0064142263,0.00019088003],"about_ca_topic_score_codex":0.000007240218,"about_ca_topic_score_gemma":0.0000049798014,"teacher_disagreement_score":0.8328173,"about_ca_system_score_codex":0.0003011543,"about_ca_system_score_gemma":0.000101699305,"threshold_uncertainty_score":0.35016757},"labels":[],"label_agreement":null},{"id":"W3115405644","doi":"10.1515/ijb-2021-0072","title":"Increasing the efficiency of randomized trial estimates via linear adjustment for a prognostic score","year":2021,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":36,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute on Aging; Canadian Institutes of Health Research","keywords":"Covariate; Estimator; Statistics; Econometrics; Variance (accounting); Sample size determination; Mathematics; Standard error; Delta method","score_opus":0.09452364034429618,"score_gpt":0.4134439329947414,"score_spread":0.31892029265044525,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3115405644","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09994035,0.0003285107,0.89702654,0.0010867554,0.0006680227,0.0008003379,0.00008058122,0.000015503056,0.000053393185],"genre_scores_gemma":[0.5988559,0.00013716253,0.4003884,0.00013109579,0.00039675354,0.000033558084,0.000011218728,0.0000190744,0.000026832944],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9979863,0.0002136799,0.0009810514,0.00008429911,0.000608883,0.00012579677],"domain_scores_gemma":[0.9811956,0.015570685,0.0011922912,0.00016289401,0.0018456433,0.000032863503],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0022759861,0.00012570345,0.00045732042,0.0000693415,0.00006231118,0.000037955055,0.0005165164,0.000038979786,0.00002817915],"category_scores_gemma":[0.0287474,0.0000659788,0.00021631332,0.00009180122,0.00025007632,0.000063231855,0.00009386146,0.00017730088,7.111441e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.40309605,0.0014019522,0.00011256617,0.00020834625,0.0020984015,0.000119818884,0.002251828,0.0004126198,0.01938669,0.548324,0.002418753,0.020169003],"study_design_scores_gemma":[0.116809525,0.0004223969,0.000020693162,0.00043175468,0.00056804676,0.0004413144,0.00020473098,0.012154571,0.042473402,0.8262741,0.00008211527,0.00011734062],"about_ca_topic_score_codex":0.000010571966,"about_ca_topic_score_gemma":0.000004981918,"teacher_disagreement_score":0.49891555,"about_ca_system_score_codex":0.00006847544,"about_ca_system_score_gemma":0.00022745368,"threshold_uncertainty_score":0.9794339},"labels":[],"label_agreement":null},{"id":"W3137145701","doi":"10.1515/ijb-2020-0130","title":"Bayesian approaches to variable selection: a comparative study from practical perspectives","year":2021,"lang":"en","type":"review","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":50,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Public Health Ontario; University of Toronto; Queen's University","funders":"Canadian Institutes of Health Research","keywords":"Computer science; Prior probability; Bayesian probability; Feature selection; Selection (genetic algorithm); Machine learning; Variable (mathematics); Bayesian inference; Artificial intelligence; Inference; Data mining; Mathematics","score_opus":0.4570566698120438,"score_gpt":0.5017153012643956,"score_spread":0.04465863145235177,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3137145701","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0000017336575,0.20754799,0.7899249,0.0004045134,0.00068593665,0.0004666677,0.00054982986,0.000010354994,0.00040804062],"genre_scores_gemma":[0.00007297243,0.15047033,0.8482039,0.00005066552,0.0010442212,0.000025256526,0.000018277684,0.00003693989,0.000077455814],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9956653,0.001326719,0.0013468735,0.0003466864,0.001095157,0.00021926877],"domain_scores_gemma":[0.98459387,0.012566088,0.0012937728,0.0002755192,0.0010985025,0.00017225093],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0013119129,0.00039414808,0.0017571403,0.00021144959,0.00011461643,0.00036961265,0.0008862204,0.00013332532,0.00057125615],"category_scores_gemma":[0.006691642,0.00024724312,0.00026494858,0.00040766218,0.00011199052,0.00009451154,0.00023833271,0.0009889829,0.000019973653],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00007707371,0.0013274422,0.0000057695615,0.00024136808,0.0037046238,0.0003477619,0.0042607444,0.000002506846,5.038197e-7,0.7788313,0.012162632,0.19903827],"study_design_scores_gemma":[0.00065111,0.0009971447,0.000013774672,0.0069586434,0.004936652,0.0014757683,0.015052368,0.00047546593,0.0000052678206,0.6249018,0.34379733,0.00073471107],"about_ca_topic_score_codex":0.000035595396,"about_ca_topic_score_gemma":0.000023605615,"teacher_disagreement_score":0.3316347,"about_ca_system_score_codex":0.0003632984,"about_ca_system_score_gemma":0.0010524229,"threshold_uncertainty_score":0.999998},"labels":[],"label_agreement":null},{"id":"W3193794660","doi":"10.1515/ijb-2020-0144","title":"Quantifying the extent of visit irregularity in longitudinal data","year":2021,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Institute of Health Services and Policy Research; Institute for Clinical Evaluative Sciences; University of Toronto; SickKids Foundation; Hospital for Sick Children; Public Health Ontario","funders":"","keywords":"Bin; Observational study; Statistics; Longitudinal data; Sample size determination; Longitudinal study; Medicine; Repeated measures design; Mathematics; Demography; Computer science; Data mining","score_opus":0.3671022536769504,"score_gpt":0.48595958555317914,"score_spread":0.11885733187622877,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3193794660","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.18930541,0.0007500146,0.800524,0.007584577,0.00091342686,0.00016379877,0.0004102841,0.00001791975,0.00033059038],"genre_scores_gemma":[0.89578986,0.00026098744,0.10367376,0.00010604498,0.00011533291,7.2298656e-7,0.000013392894,0.000009780323,0.000030093415],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9983418,0.000120526434,0.00066864054,0.00010277923,0.00065956626,0.000106685045],"domain_scores_gemma":[0.9963416,0.0017044141,0.0006404115,0.00046125418,0.0008279701,0.00002434104],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013417369,0.00009019617,0.00020238591,0.000069745074,0.00003466842,0.000048388076,0.0013506901,0.00003276498,0.000064215776],"category_scores_gemma":[0.0033310226,0.000053385003,0.00004979569,0.000128769,0.00013500474,0.00015868926,0.00045004993,0.0003206368,0.0000014331173],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00019653297,0.0005080045,0.011631263,0.000068957095,0.00043468102,0.0008207782,0.00093380944,0.00014017896,0.012495126,0.94197613,0.008337765,0.022456802],"study_design_scores_gemma":[0.00057243905,0.00009883024,0.020871291,0.0005613747,0.00012421861,0.0010400115,0.0010891979,0.0036589452,0.028580388,0.9417871,0.0014478515,0.00016833056],"about_ca_topic_score_codex":0.000028147659,"about_ca_topic_score_gemma":0.00021320155,"teacher_disagreement_score":0.7064845,"about_ca_system_score_codex":0.0000753376,"about_ca_system_score_gemma":0.00015903653,"threshold_uncertainty_score":0.3987786},"labels":[],"label_agreement":null},{"id":"W3197649143","doi":"10.1515/ijb-2020-0046","title":"An improved method for analysis of interrupted time series (ITS) data: accounting for patient heterogeneity using weighted analysis","year":2021,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Blood Pressure and Hypertension Studies","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; McMaster University; Children's Hospital of Eastern Ontario; Impact; St. Michael's Hospital; St. Joseph’s Healthcare Hamilton; University of Ottawa","funders":"","keywords":"Interrupted Time Series Analysis; Time series; Statistics; Series (stratigraphy); Econometrics; Interrupted time series; Computer science; Mathematics; Accounting; Medicine; Economics","score_opus":0.078531408221474,"score_gpt":0.39894760312035393,"score_spread":0.32041619489887996,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3197649143","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.25551677,0.0012597447,0.7363287,0.001665127,0.0003925071,0.00022539987,0.004594634,0.000007835067,0.000009241259],"genre_scores_gemma":[0.5976403,0.00008733355,0.40024418,0.00093089155,0.00024205122,0.0000025971042,0.0008048626,0.000015376523,0.00003242757],"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984949,0.000082192535,0.0007433257,0.00017802979,0.00038585928,0.00011569455],"domain_scores_gemma":[0.9939112,0.00068794686,0.0008071085,0.00031053717,0.004231225,0.00005200167],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00073983974,0.00011624184,0.0007378336,0.0003806132,0.00007855511,0.0000528628,0.0003282594,0.000045090455,0.000040311832],"category_scores_gemma":[0.00067116943,0.00007864759,0.00033897633,0.0004758029,0.00004280963,0.00016412292,0.00014788235,0.000092976814,1.8909272e-7],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.005279258,0.0011653834,0.012458041,0.00017207632,0.35490996,0.00013745256,0.0021090822,0.003063316,0.5846335,0.00056925596,0.0040511168,0.03145154],"study_design_scores_gemma":[0.0012633561,0.00042734508,0.0033237434,0.000066779474,0.0944432,0.00013823828,0.0010975695,0.8381983,0.057788327,0.0002072184,0.0029018892,0.00014403599],"about_ca_topic_score_codex":0.000030666357,"about_ca_topic_score_gemma":0.000074986165,"teacher_disagreement_score":0.835135,"about_ca_system_score_codex":0.000024805,"about_ca_system_score_gemma":0.00013313613,"threshold_uncertainty_score":0.32071555},"labels":[],"label_agreement":null},{"id":"W4294310517","doi":"10.1515/ijb-2022-0056","title":"Causal inference for oncology: past developments and current challenges","year":2022,"lang":"en","type":"review","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"Fonds de Recherche du Québec - Santé","keywords":"Causal inference; Inference; Current (fluid); Medical physics; Medicine; Econometrics; Data science; Computer science; Oncology; Mathematics; Artificial intelligence; Physics","score_opus":0.48275920743690715,"score_gpt":0.5532970978116655,"score_spread":0.07053789037475833,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4294310517","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0000016403741,0.93425685,0.06281343,0.0003185928,0.0014580927,0.00047155152,0.0004918167,0.000022363865,0.00016566452],"genre_scores_gemma":[0.000014154844,0.93131566,0.06807915,0.000037820955,0.00037144922,0.00007243268,0.00004035722,0.00003810036,0.000030872776],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99782133,0.00017188133,0.0010656506,0.00016864175,0.0005939136,0.00017857584],"domain_scores_gemma":[0.9903691,0.0069591138,0.0019471899,0.00015959908,0.0004942515,0.000070793816],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011247034,0.0002865315,0.000903183,0.00023268205,0.00009164563,0.00005445914,0.0009981262,0.00010373414,0.000048853537],"category_scores_gemma":[0.002651666,0.00018809573,0.00015383298,0.00006415095,0.00012425958,0.00008353592,0.00038272783,0.00069558836,0.0000018314092],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00001785379,0.000072047806,6.528715e-7,0.0010795246,0.00029136418,0.000027528993,0.00022981084,1.5304565e-7,2.1767487e-7,0.094655834,0.0025565785,0.90106845],"study_design_scores_gemma":[0.0001541694,0.0001955814,0.0000013129084,0.0020796324,0.00031306685,0.00031245875,0.00008080421,0.0000042389847,0.0000016997569,0.1792551,0.81744474,0.00015718631],"about_ca_topic_score_codex":9.895064e-7,"about_ca_topic_score_gemma":0.000006645366,"teacher_disagreement_score":0.9009113,"about_ca_system_score_codex":0.00053595076,"about_ca_system_score_gemma":0.00069807726,"threshold_uncertainty_score":0.76703215},"labels":[],"label_agreement":null},{"id":"W4307134400","doi":"10.1515/ijb-2022-0010","title":"A copula-based set-variant association test for bivariate continuous, binary or mixed phenotypes","year":2022,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal; McGill University","funders":"Medical Research Council; Natural Sciences and Engineering Research Council of Canada; Wellcome Trust","keywords":"Bivariate analysis; Multivariate statistics; Copula (linguistics); Type I and type II errors; Genome-wide association study; Multivariate normal distribution; Statistics; Multivariate analysis; Genetic association; Mathematics; Computer science; Econometrics; Biology; Genetics; Single-nucleotide polymorphism; Gene","score_opus":0.01717379251875667,"score_gpt":0.2889192896215963,"score_spread":0.27174549710283963,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4307134400","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.643559,0.0006984588,0.30897552,0.032013707,0.0071379673,0.00074151903,0.0066360077,0.000018521494,0.00021932216],"genre_scores_gemma":[0.975575,0.00006653759,0.020691514,0.0018282856,0.00056391966,0.000023081553,0.00037439464,0.00001858486,0.0008587185],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.998772,0.00015438763,0.00046955264,0.00011819058,0.00032263753,0.00016321968],"domain_scores_gemma":[0.99753726,0.0008093953,0.00088454026,0.00011444802,0.0006153875,0.0000389607],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013567853,0.000098251985,0.00017360356,0.00006380538,0.00018093082,0.00003165603,0.00045592096,0.00005992151,0.00007776633],"category_scores_gemma":[0.0034614676,0.0000716307,0.000108122316,0.00006635704,0.000029796056,0.0000029837531,0.00011164477,0.00013593056,0.000002202846],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.004150429,0.0010846278,0.065113954,0.000029708659,0.0024614965,0.00014334252,0.0004630246,0.030435277,0.14770651,0.0034881197,0.7340437,0.010879764],"study_design_scores_gemma":[0.012649214,0.010015475,0.10096107,0.00006409889,0.0008302998,0.00074171997,0.0024021976,0.04913305,0.010746241,0.011599197,0.799834,0.0010234499],"about_ca_topic_score_codex":0.000027598267,"about_ca_topic_score_gemma":0.000032216958,"teacher_disagreement_score":0.332016,"about_ca_system_score_codex":0.00015602329,"about_ca_system_score_gemma":0.00033271994,"threshold_uncertainty_score":0.414395},"labels":[],"label_agreement":null},{"id":"W4310956147","doi":"10.1515/ijb-2022-0017","title":"Statistics, philosophy, and health: the SMAC 2021 webconference","year":2022,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Philosophy and History of Science","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Causality (physics); Table (database); Point (geometry); Computer science; Epistemology; Statistics; Library science; Mathematics; Philosophy; Data mining","score_opus":0.08710155087595003,"score_gpt":0.2798330896389365,"score_spread":0.19273153876298646,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4310956147","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.011342678,0.009024031,0.015240432,0.8490292,0.027176883,0.000656529,0.0108197015,0.000036008754,0.076674566],"genre_scores_gemma":[0.99187857,0.00028530796,0.0008414636,0.004896661,0.001230109,0.0000025659872,0.000010919948,0.0000066802654,0.00084770954],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99876565,0.00008476794,0.00032182856,0.00007653863,0.00064785406,0.00010334932],"domain_scores_gemma":[0.99889827,0.00024216289,0.0004206499,0.000089523935,0.00030019306,0.000049220736],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00078147586,0.000073580144,0.00010591167,0.00007059444,0.0009233083,0.00015998419,0.00076669146,0.0000054319075,0.0015074251],"category_scores_gemma":[0.00005430659,0.00004434729,0.00003145259,0.000022899232,0.0006636969,0.00009594701,0.00015887895,0.00029516616,0.000009904521],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000031873446,0.000033326753,0.000019238336,0.0000030816452,0.00004107451,0.000023623614,0.008630532,0.000023366258,0.000010677919,0.92567366,0.054320246,0.011189296],"study_design_scores_gemma":[0.00024497218,0.00033198006,0.00025662626,0.000018029834,0.000016928647,0.00015347089,0.002769605,0.0003247491,0.0000045111274,0.41108987,0.58470976,0.000079492485],"about_ca_topic_score_codex":0.0001413839,"about_ca_topic_score_gemma":0.0001271916,"teacher_disagreement_score":0.9805359,"about_ca_system_score_codex":0.00009557226,"about_ca_system_score_gemma":0.00020911134,"threshold_uncertainty_score":0.9994053},"labels":[],"label_agreement":null},{"id":"W4375864930","doi":"10.1515/ijb-2021-0091","title":"Error analysis of the PacBio sequencing CCS reads","year":2023,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Genomics and Phylogenetic Studies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Nanopore sequencing; Computer science; Hybrid genome assembly; Word error rate; Probabilistic logic; Sequence assembly; DNA sequencing; Process (computing); k-mer; Error detection and correction; Binomial distribution; Algorithm; Computational biology; Shotgun sequencing; Data mining; Biology; Mathematics; Statistics; Artificial intelligence; Genetics; DNA","score_opus":0.02603351602590837,"score_gpt":0.29465399206466025,"score_spread":0.2686204760387519,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4375864930","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99660915,0.00021139541,0.0009341303,0.0011095001,0.0006343796,0.000034662327,0.00027052723,7.965444e-7,0.00019543621],"genre_scores_gemma":[0.99843055,0.00043737108,0.0005213816,0.00021416675,0.00017149972,5.872965e-7,0.000018823705,0.0000060022435,0.0001996461],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9992635,0.000040822622,0.00028881626,0.00006486756,0.00026152484,0.00008046969],"domain_scores_gemma":[0.9990644,0.00007701266,0.00032661922,0.00015436322,0.00035938827,0.00001821843],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003443974,0.00006354549,0.00011723167,0.00008424145,0.000047206213,0.000014206562,0.0004964757,0.00003192609,0.0000063049692],"category_scores_gemma":[0.00025438648,0.00003616191,0.00014018938,0.00022030683,0.0000978135,5.7986733e-7,0.00015705475,0.00006683644,0.0000015139942],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009988674,0.000028245919,0.017991312,0.000005903963,0.005923603,0.000014000712,0.00068068795,0.008545506,0.9521201,0.0018775502,0.010369023,0.002344161],"study_design_scores_gemma":[0.0014658713,0.0006300653,0.52346426,0.000088566776,0.0037074524,0.00018849324,0.0040480043,0.00867216,0.40159974,0.007423462,0.048212778,0.00049913296],"about_ca_topic_score_codex":0.000024359484,"about_ca_topic_score_gemma":0.000055661025,"teacher_disagreement_score":0.5505204,"about_ca_system_score_codex":0.000019991297,"about_ca_system_score_gemma":0.00009089436,"threshold_uncertainty_score":0.14746399},"labels":[],"label_agreement":null},{"id":"W4376112662","doi":"10.1515/ijb-2022-0040","title":"Exact correction factor for estimating the OR in the presence of sparse data with a zero cell in 2 × 2 tables","year":2023,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Impact; McMaster University; Population Health Research Institute","funders":"","keywords":"Zero (linguistics); Statistics; Mathematics; Applied mathematics; Computer science; Algorithm","score_opus":0.14832876308927137,"score_gpt":0.4169003802549265,"score_spread":0.26857161716565514,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4376112662","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.02054098,0.000013709594,0.9767212,0.0010426542,0.00057775097,0.0002585638,0.0007831065,0.0000040720724,0.000057937712],"genre_scores_gemma":[0.47733513,0.0000336672,0.52236885,0.00007036172,0.00011118785,0.0000075183148,0.0000118171165,0.000010831858,0.000050649647],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99852955,0.00019400902,0.00053398317,0.00009683481,0.00051224127,0.00013338262],"domain_scores_gemma":[0.9814975,0.01744425,0.0005299179,0.0002505295,0.0002597331,0.000018055529],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0018871751,0.000091436435,0.00017929033,0.000096337775,0.000047802154,0.000069997375,0.001153618,0.000025341513,0.000031252395],"category_scores_gemma":[0.0092731295,0.00003740686,0.000022727403,0.00025820234,0.00011560332,0.000098644334,0.00010203398,0.00022676842,0.0000010370917],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0061600003,0.0016147147,0.011298787,0.0009034007,0.00062620704,0.0009078627,0.033924825,0.006554232,0.003224578,0.27083117,0.11692082,0.5470334],"study_design_scores_gemma":[0.001335122,0.00051855756,0.008976883,0.00082858594,0.00009488571,0.00022965817,0.0033031239,0.49414665,0.00096678076,0.48894346,0.0004912431,0.00016506613],"about_ca_topic_score_codex":0.000088315595,"about_ca_topic_score_gemma":0.00025663085,"teacher_disagreement_score":0.5468683,"about_ca_system_score_codex":0.000031362426,"about_ca_system_score_gemma":0.0001412162,"threshold_uncertainty_score":0.9990722},"labels":[],"label_agreement":null},{"id":"W4376637825","doi":"10.1515/ijb-2022-0073","title":"Bayesian inference for optimal dynamic treatment regimes in practice","year":2023,"lang":"en","type":"review","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Inference; Bayesian inference; Bayesian probability; Frequentist inference; Biostatistics; Econometrics; Estimator; Machine learning; Computer science; Artificial intelligence; Mathematics; Statistics; Medicine","score_opus":0.22417419034281136,"score_gpt":0.5431292370587046,"score_spread":0.31895504671589325,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4376637825","genre_codex":"methods","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0000017444507,0.41022655,0.5862666,0.0006756311,0.0008022691,0.0010078473,0.0007873179,0.00005895745,0.0001731272],"genre_scores_gemma":[0.000020334639,0.7893362,0.20991872,0.00003389114,0.0001736952,0.00006325346,0.00004406126,0.00006096463,0.00034887993],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9974878,0.00018031223,0.0013939728,0.00018022532,0.00053541845,0.00022229455],"domain_scores_gemma":[0.98258305,0.0140299415,0.002298841,0.00026820294,0.0007634849,0.00005649651],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0009862275,0.00034191035,0.0010396782,0.00041408846,0.000039727747,0.00010157255,0.0010345423,0.00016027532,0.000011403931],"category_scores_gemma":[0.008670338,0.00021249938,0.00030293787,0.00019165814,0.00009512311,0.00017752942,0.00011299426,0.00043622087,0.000010382513],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000101521626,0.00022148252,5.034768e-7,0.001112301,0.0011158044,0.00038167374,0.00033929382,0.00003707124,9.670149e-7,0.13737918,0.0019989966,0.8573112],"study_design_scores_gemma":[0.00047999434,0.0006302794,4.6917654e-7,0.011685474,0.0013732146,0.0006652835,0.0002798653,0.00085759006,0.000009326241,0.2821943,0.7014732,0.00035101056],"about_ca_topic_score_codex":0.000016340453,"about_ca_topic_score_gemma":0.000045598317,"teacher_disagreement_score":0.8569602,"about_ca_system_score_codex":0.0009664516,"about_ca_system_score_gemma":0.0006871058,"threshold_uncertainty_score":0.99968004},"labels":[],"label_agreement":null},{"id":"W4394615219","doi":"10.1515/ijb-2023-0102","title":"Ensemble learning methods of inference for spatially stratified infectious disease systems","year":2024,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Animal Disease Management and Epidemiology","field":"Agricultural and Biological Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Winnipeg; University of Calgary","funders":"","keywords":"Markov chain Monte Carlo; Covariate; Computer science; Bayesian inference; Inference; Bayesian probability; Ensemble learning; Population; Boosting (machine learning); Random forest; Machine learning; Markov chain; Artificial intelligence; Statistical inference; Bayes' theorem; Statistics; Econometrics; Mathematics","score_opus":0.049532261754997794,"score_gpt":0.3602157903403812,"score_spread":0.3106835285853834,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394615219","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.54014015,0.0023433706,0.44735947,0.005114776,0.0032725686,0.0003888829,0.00038234005,0.000042947497,0.00095548434],"genre_scores_gemma":[0.9975589,0.00019744474,0.0015898199,0.00008045541,0.0004064096,0.0000034909995,0.000027905538,9.469125e-7,0.00013463065],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9991678,0.00014474911,0.00035107625,0.00007292768,0.00017954322,0.000083899315],"domain_scores_gemma":[0.99718314,0.0022077619,0.00023532921,0.000019187753,0.00030573303,0.000048829963],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00079157925,0.00006495252,0.00012873397,0.000023351618,0.000047275367,0.0000848356,0.00026512388,0.00002274129,0.000053674627],"category_scores_gemma":[0.001127789,0.000022959348,0.000092437054,0.00006635869,0.00004851583,0.0000617627,0.000042425578,0.000095760835,0.0000028125573],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00074319186,0.00013534768,0.007292328,0.00016080072,0.00066919724,0.00007678827,0.00027620947,0.0059702992,0.08099093,0.32366458,0.0072314003,0.5727889],"study_design_scores_gemma":[0.0009728674,0.0043395786,0.20459192,0.0011549942,0.0010618444,0.0000884883,0.0018254365,0.36168554,0.0020338593,0.23233998,0.18916205,0.0007434451],"about_ca_topic_score_codex":0.00005890572,"about_ca_topic_score_gemma":0.00002010455,"teacher_disagreement_score":0.57204545,"about_ca_system_score_codex":0.000020867306,"about_ca_system_score_gemma":0.000030316734,"threshold_uncertainty_score":0.13501503},"labels":[],"label_agreement":null},{"id":"W4394618888","doi":"10.1515/ijb-2023-0121","title":"The survival function NPMLE for combined right-censored and length-biased right-censored failure time data: properties and applications","year":2024,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"McGill University; University of Regina","funders":"","keywords":"Estimator; Survival function; Kaplan–Meier estimator; Cohort; Nonparametric statistics; Statistics; Function (biology); Mathematics; Survival analysis; Econometrics; Maximum likelihood; Cohort effect","score_opus":0.07410926737206686,"score_gpt":0.339667379602726,"score_spread":0.26555811223065917,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394618888","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0073024784,0.0010677126,0.9780827,0.009623074,0.0011957587,0.0006744563,0.0016934907,0.00004297148,0.00031734086],"genre_scores_gemma":[0.5722491,0.0011011538,0.4214779,0.00040385887,0.0021351606,0.00008611127,0.00017889267,0.000099275305,0.0022685472],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985967,0.00012954243,0.00052593974,0.00017980202,0.0004182296,0.00014976854],"domain_scores_gemma":[0.9930921,0.0057726735,0.00024451767,0.00024194807,0.00057886326,0.0000698491],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011945204,0.00014628231,0.00021230835,0.00006254044,0.00026203968,0.0004155736,0.0005149014,0.00005166669,0.00004447536],"category_scores_gemma":[0.0019483696,0.00007492524,0.00004037962,0.0000733151,0.00028648763,0.00013221851,0.00012810789,0.00022365198,0.0000076022034],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003738375,0.000063492036,0.000018134871,0.0000801213,0.0004497719,0.000011222247,0.0001716048,0.0000019867964,0.0021873526,0.9490372,0.023164006,0.024441306],"study_design_scores_gemma":[0.0009861484,0.0003403396,0.0003360474,0.0002040069,0.00040565635,0.00012625038,0.00040096595,0.04603768,0.00081743445,0.8623783,0.087765746,0.00020141645],"about_ca_topic_score_codex":0.000008823595,"about_ca_topic_score_gemma":0.000016554704,"teacher_disagreement_score":0.56494665,"about_ca_system_score_codex":0.00004161675,"about_ca_system_score_gemma":0.0000947341,"threshold_uncertainty_score":0.4007386},"labels":[],"label_agreement":null},{"id":"W4395117975","doi":"10.1515/ijb-2023-0056","title":"Random forests for survival data: which methods work best and under what conditions?","year":2024,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Random forest; Statistics; Work (physics); Mathematics; Econometrics; Computer science; Engineering; Machine learning","score_opus":0.21037535950266203,"score_gpt":0.5072925026780893,"score_spread":0.2969171431754273,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4395117975","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0023504773,0.0011519534,0.9869739,0.0050087105,0.0034999182,0.00015034952,0.0007623928,0.000010790249,0.00009152091],"genre_scores_gemma":[0.075532444,0.0009083481,0.9223636,0.00022709393,0.0007187916,0.0000064231467,0.00004887802,0.000027245525,0.00016720641],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99854594,0.00022357724,0.0005453593,0.00015125616,0.00039825513,0.00013561311],"domain_scores_gemma":[0.9769316,0.021876708,0.00021466997,0.00019881604,0.0007048711,0.00007333942],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027671775,0.00012481354,0.00025508145,0.000093374074,0.000076546705,0.0005928311,0.00059071765,0.000049701433,0.00008437526],"category_scores_gemma":[0.007175998,0.00007822427,0.00004998865,0.00012393104,0.00013926518,0.00028488156,0.00015202134,0.00023232456,0.000004243193],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002749417,0.00006372184,0.00007755187,0.00007204786,0.00057447574,0.000029529945,0.00024735896,0.000025229663,0.00016310479,0.87392944,0.016042504,0.108500116],"study_design_scores_gemma":[0.00073994225,0.000112061025,0.00045456222,0.0005019292,0.00027077805,0.00015600104,0.0005200736,0.011339467,0.00014997978,0.98006535,0.0055760858,0.00011380079],"about_ca_topic_score_codex":0.0000055605883,"about_ca_topic_score_gemma":0.000028748931,"teacher_disagreement_score":0.10838631,"about_ca_system_score_codex":0.000040561663,"about_ca_system_score_gemma":0.00013214063,"threshold_uncertainty_score":0.8590858},"labels":[],"label_agreement":null},{"id":"W4401517718","doi":"10.1515/ijb-2023-0116","title":"History-restricted marginal structural model and latent class growth analysis of treatment trajectories for a time-dependent outcome","year":2024,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University Health Centre; Université de Montréal; McGill University; Institut National de Santé Publique du Québec; Université Laval","funders":"Canadian Institutes of Health Research; Université Laval","keywords":"Econometrics; Marginal structural model; Outcome (game theory); Class (philosophy); Latent class model; Structural equation modeling; Statistics; Marginal model; Mathematics; Computer science; Causal inference; Mathematical economics; Regression analysis; Artificial intelligence","score_opus":0.10769438102764733,"score_gpt":0.39364910278080345,"score_spread":0.2859547217531561,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4401517718","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.33597946,0.00057241344,0.6600345,0.0009066689,0.00056755036,0.000293284,0.001509899,0.000057164572,0.00007907886],"genre_scores_gemma":[0.9318358,0.00014605261,0.0674944,0.000031558266,0.00007327502,0.000006291657,0.000019728444,0.000017406104,0.0003755408],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9987711,0.000030475423,0.0006114189,0.00010506512,0.00037749493,0.000104458966],"domain_scores_gemma":[0.9978052,0.0010517333,0.00039758332,0.000097352044,0.0006036549,0.00004448863],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00021421214,0.00014206165,0.0003507646,0.00034891116,0.000023947869,0.00003983835,0.0002562714,0.000043103617,0.000037267117],"category_scores_gemma":[0.0003876831,0.0000877755,0.00016109661,0.00010733394,0.00010109524,0.00010832697,0.00003610197,0.00011399332,3.9072933e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009270431,0.00016667685,0.0029754806,0.00016717073,0.01074213,0.00012700389,0.0030189867,0.0047080307,0.016148042,0.95123464,0.004325324,0.0054594814],"study_design_scores_gemma":[0.00070480007,0.00075875747,0.0013437972,0.00009471222,0.003682103,0.00011531715,0.00010156778,0.48307687,0.0024136782,0.5071725,0.00032096007,0.00021489162],"about_ca_topic_score_codex":0.00002695437,"about_ca_topic_score_gemma":0.000024030822,"teacher_disagreement_score":0.5958563,"about_ca_system_score_codex":0.0007628059,"about_ca_system_score_gemma":0.00015541996,"threshold_uncertainty_score":0.3579381},"labels":[],"label_agreement":null},{"id":"W4410591660","doi":"10.1515/ijb-2024-0005","title":"Guidance on individualized treatment rule estimation in high dimensions","year":2025,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Interpretability; Covariate; Observational study; Estimator; Computer science; Machine learning; Mathematics; Statistics","score_opus":0.057206199167839276,"score_gpt":0.41183681128257044,"score_spread":0.3546306121147312,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4410591660","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2398816,0.00010589606,0.74801546,0.0068378616,0.0019368448,0.0002381362,0.00034120146,0.000015967746,0.002627017],"genre_scores_gemma":[0.5483385,0.00007216862,0.45087773,0.00044599894,0.00005068298,0.0000048885695,0.0000050942754,0.000006731907,0.00019823232],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99868554,0.0001476484,0.00058367464,0.00008582765,0.00039060388,0.000106699925],"domain_scores_gemma":[0.99525666,0.0039524185,0.0003279632,0.0001357256,0.00029486755,0.00003238525],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005775156,0.00010769987,0.00021450281,0.00018571697,0.00004512961,0.000049551632,0.00033514996,0.00003692466,0.000065497254],"category_scores_gemma":[0.0044112653,0.000066168504,0.0000472425,0.00012923554,0.00006627458,0.000035079865,0.000042062275,0.00014629519,0.000008323899],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016221342,0.00025807033,0.0005572353,0.000007023354,0.00013229335,0.000048310932,0.00021662933,0.00025472642,0.0002276497,0.93621886,0.0032991797,0.058617804],"study_design_scores_gemma":[0.0012155995,0.00021191705,0.009994183,0.00029257566,0.000059316855,0.000016871232,0.000068156725,0.004228194,0.0020897833,0.9809647,0.0007862831,0.00007240673],"about_ca_topic_score_codex":0.000044955326,"about_ca_topic_score_gemma":0.000015017379,"teacher_disagreement_score":0.30845687,"about_ca_system_score_codex":0.00021653307,"about_ca_system_score_gemma":0.000126118,"threshold_uncertainty_score":0.5281015},"labels":[],"label_agreement":null},{"id":"W4416438862","doi":"10.1515/ijb-2024-0011","title":"Post-shrinkage strategies for nonlinear semiparametric regression models in low and high-dimensional settings","year":2025,"lang":"en","type":"article","venue":"The International Journal of Biostatistics","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Brock University","funders":"Türkiye Bilimsel ve Teknolojik Araştırma Kurumu","keywords":"Estimator; Parametric statistics; Semiparametric regression; Nonlinear system; Semiparametric model; Regression; Estimation; Parametric model; Nonlinear regression; Breast cancer","score_opus":0.034606502573104284,"score_gpt":0.36961705616745444,"score_spread":0.33501055359435017,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416438862","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1765745,0.00017478013,0.8185351,0.0032165344,0.000628914,0.00017607423,0.00052296725,0.000007303332,0.00016382737],"genre_scores_gemma":[0.43543226,0.00005670274,0.5637537,0.0005672759,0.000085047635,0.0000032845278,0.00000824205,0.000008898967,0.000084570325],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99870574,0.0000807116,0.0005916942,0.00011383302,0.00037353026,0.0001344585],"domain_scores_gemma":[0.9932617,0.0054028453,0.00035339282,0.0000955635,0.0008445079,0.000041995594],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009157138,0.000120665754,0.00024377834,0.00025951402,0.000052667816,0.000100135716,0.0003223407,0.000058773825,0.000015942189],"category_scores_gemma":[0.004001429,0.000075984855,0.0000457797,0.00013965124,0.00010151762,0.00011087992,0.00009415408,0.00023900809,5.343811e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00037211657,0.00014209271,0.00007537905,0.000117548254,0.00010705639,0.000042836604,0.00027380144,0.00045705005,0.0009925687,0.96615034,0.0028355469,0.02843365],"study_design_scores_gemma":[0.00076638843,0.00012189076,0.0005316429,0.0005873847,0.000039284067,0.000038687624,0.00029436793,0.093952395,0.00082722254,0.9026818,0.00007894861,0.00007994],"about_ca_topic_score_codex":0.000025320183,"about_ca_topic_score_gemma":0.000009828382,"teacher_disagreement_score":0.25885776,"about_ca_system_score_codex":0.000058110225,"about_ca_system_score_gemma":0.00020924445,"threshold_uncertainty_score":0.47903734},"labels":[],"label_agreement":null}]}