{"meta":{"query_hash":"0e940ce2e494","filters":{"venue":"Statistics in Biosciences"},"cohort_total":42,"direct_labels_cover":0,"predictions_cover":42,"exported":42,"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/0e940ce2e494","api":"https://metacan.xera.ac/api/v1/cohort?venue=Statistics+in+Biosciences"},"results":[{"id":"W2015341825","doi":"10.1007/s12561-010-9031-0","title":"Cox Regression with Covariates Missing Not at Random","year":2010,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":11,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University; BC Centre for Disease Control; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Public Health Agency","keywords":"Covariate; Proportional hazards model; Missing data; Estimation; Econometrics; Statistics; Biostatistics; Computer science; Regression; Event (particle physics); Regression analysis; Data mining; Mathematics; Public health; Engineering; Medicine","score_opus":0.051099391357166526,"score_gpt":0.38576229698868225,"score_spread":0.33466290563151574,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2015341825","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":[0.029929463,0.000008381335,0.9672125,0.00030736998,0.0005058732,0.00018089623,0.00020005737,0.00003795256,0.001617528],"genre_scores_gemma":[0.21327646,0.0000075537355,0.78648144,0.00006747066,0.000034868735,0.000008211076,0.0000031767984,0.000009110947,0.00011170035],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99843884,0.000120550474,0.000325579,0.00034541616,0.0004380147,0.00033162755],"domain_scores_gemma":[0.9951516,0.004236876,0.00016724737,0.00023898408,0.00008636269,0.000118925185],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010006693,0.00016576218,0.00028559536,0.00009581242,0.00025619433,0.000109519264,0.00026816325,0.000076828146,0.00037866147],"category_scores_gemma":[0.0046897214,0.000101689046,0.000016116095,0.0003099711,0.00075013057,0.00007863245,0.0000712969,0.00027297795,0.000007869502],"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":[0.0001444534,0.00007135397,0.0064076935,0.000055512875,0.0000038863914,0.000052856696,0.0003871708,2.996395e-7,0.011578592,0.95493996,0.0007767067,0.025581522],"study_design_scores_gemma":[0.0009050361,0.00013853254,0.008478395,0.00014212551,0.000018420746,0.00002860369,0.00010358236,0.0055578984,0.0070977025,0.97688663,0.0003735409,0.0002695094],"about_ca_topic_score_codex":0.000059200338,"about_ca_topic_score_gemma":0.0005558471,"teacher_disagreement_score":0.18334699,"about_ca_system_score_codex":0.000024792838,"about_ca_system_score_gemma":0.00010079418,"threshold_uncertainty_score":0.56143737},"labels":[],"label_agreement":null},{"id":"W2018472295","doi":"10.1007/s12561-013-9103-z","title":"Q-Learning: Flexible Learning About Useful Utilities","year":2013,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":110,"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","funders":"","keywords":"Observational study; Machine learning; Biostatistics; Computer science; Covariate; Linear model; Artificial intelligence; Econometrics; Mathematical optimization; Mathematics; Statistics; Medicine","score_opus":0.1224448115109042,"score_gpt":0.40824360456375525,"score_spread":0.28579879305285105,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2018472295","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":[0.46557373,0.00020748773,0.47646892,0.00030928355,0.0005605749,0.0010395537,0.000058150483,0.001892074,0.053890217],"genre_scores_gemma":[0.7288471,0.00014434413,0.2654428,0.00004480861,0.000027035383,0.00006616394,0.0000051601814,0.000017811608,0.0054048058],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99821925,0.00012895413,0.00040491534,0.00034441098,0.0004105313,0.00049196085],"domain_scores_gemma":[0.9981221,0.0012484844,0.00018505,0.00019713416,0.00015907476,0.00008814776],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00046240893,0.00018834823,0.0002515057,0.00020802488,0.00022471575,0.00017042684,0.0003769468,0.00008348388,0.000736755],"category_scores_gemma":[0.0037929725,0.00016327619,0.0000250266,0.0004442045,0.0006029133,0.0004051474,0.00012464696,0.00044204723,0.000102134385],"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":[0.000005126725,0.00009315108,0.055582155,0.000115339506,0.000006497476,0.00001070485,0.0030808754,0.000103320024,0.0019193965,0.9150398,0.0041896687,0.019853987],"study_design_scores_gemma":[0.00009527897,0.00026067483,0.0100502195,0.00008440084,0.0000045694323,0.0000036479778,0.0038822666,0.0036663988,0.003100653,0.9755664,0.0030016045,0.00028391607],"about_ca_topic_score_codex":0.00020685639,"about_ca_topic_score_gemma":0.00008362322,"teacher_disagreement_score":0.26327336,"about_ca_system_score_codex":0.0000708594,"about_ca_system_score_gemma":0.00007693613,"threshold_uncertainty_score":0.8066951},"labels":[],"label_agreement":null},{"id":"W2034407073","doi":"10.1007/s12561-013-9083-z","title":"The Design of Intervention Trials Involving Recurrent and Terminal Events","year":2013,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","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":"University of Waterloo","funders":"","keywords":"Statistics; Censoring (clinical trials); Event (particle physics); Sample size determination; Markov chain; Marginal model; Econometrics; Computer science; Regression analysis; Mathematics","score_opus":0.2570574189063031,"score_gpt":0.46297195949869513,"score_spread":0.205914540592392,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2034407073","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":[0.043324832,0.00024440928,0.95532936,0.000080426056,0.00039519224,0.00045579893,0.00006080443,0.0000066322905,0.00010251658],"genre_scores_gemma":[0.34688413,0.00021162185,0.6528061,0.0000070575015,0.000012663522,0.00003247176,5.35643e-7,0.0000031221878,0.00004228307],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99820024,0.0006720387,0.00059819076,0.00014366521,0.00022635497,0.00015948869],"domain_scores_gemma":[0.98927975,0.010205301,0.00028346945,0.000102858474,0.00008881067,0.00003982582],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.004384214,0.00007664715,0.000222262,0.000053617612,0.000096902346,0.00005863322,0.00018078073,0.000026183647,0.00006287697],"category_scores_gemma":[0.022292389,0.000044010132,0.000018498105,0.0001375318,0.00033683117,0.000063791456,0.000060558305,0.00007421598,0.000002082184],"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":[0.000016036473,0.00008164675,0.001999244,0.00009347353,0.0000063133125,9.951167e-7,0.00039185697,6.03666e-7,0.0007643252,0.6463766,0.0007306839,0.34953824],"study_design_scores_gemma":[0.00014192345,0.0002119846,0.013330509,0.00013896513,0.000009636021,0.0000013157133,0.00045095402,0.0110521875,0.00029585234,0.9742912,0.000013688272,0.00006176286],"about_ca_topic_score_codex":0.00007553089,"about_ca_topic_score_gemma":0.00003226119,"teacher_disagreement_score":0.3494765,"about_ca_system_score_codex":0.000014112354,"about_ca_system_score_gemma":0.000028289,"threshold_uncertainty_score":0.98594326},"labels":[],"label_agreement":null},{"id":"W2086560295","doi":"10.1007/s12561-011-9038-1","title":"A Data Augmentation Method for Estimating the Causal Effect of Adherence to Treatment Regimens Targeting Control of an Intermediate Measure","year":2011,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Regimen; Censoring (clinical trials); Covariate; Causal inference; Medicine; Observational study; Inverse probability weighting; Hazard ratio; Proportional hazards model; Repeated measures design; Statistics; Internal medicine; Econometrics; Mathematics; Confidence interval","score_opus":0.2405471980053799,"score_gpt":0.49007497699211644,"score_spread":0.24952777898673653,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2086560295","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":[0.039464388,0.0000122649635,0.95820504,0.0000108824215,0.000121326615,0.0010177556,0.0010983993,0.000026299182,0.00004364511],"genre_scores_gemma":[0.4334019,0.0000015212735,0.566473,0.0000065760787,0.000009630744,0.00008448175,0.000014191597,0.0000057205466,0.0000029519992],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984772,0.00028647503,0.00048810773,0.00029496863,0.000248256,0.00020495566],"domain_scores_gemma":[0.99595577,0.003020081,0.0004362748,0.0004122972,0.0001299581,0.000045621353],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027472319,0.00014080074,0.00032006513,0.00008605745,0.00006749108,0.000013968777,0.0005696128,0.0000349919,0.00000904476],"category_scores_gemma":[0.005028422,0.00008686729,0.000018797997,0.00020630544,0.00022174344,0.00018274592,0.000077289864,0.00005656893,2.9032296e-7],"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":[0.001388159,0.0009024189,0.00783785,0.0012039632,0.00017466865,0.000016361788,0.041628994,0.000496136,0.16980027,0.18582039,0.0005808418,0.59014994],"study_design_scores_gemma":[0.0015237128,0.012608213,0.0007862359,0.0004417517,0.00021167399,0.0000036455872,0.0023048709,0.3572751,0.27336636,0.35111454,0.000010378775,0.000353501],"about_ca_topic_score_codex":0.00028827402,"about_ca_topic_score_gemma":0.0005450785,"teacher_disagreement_score":0.5897965,"about_ca_system_score_codex":0.000049573213,"about_ca_system_score_gemma":0.00007029888,"threshold_uncertainty_score":0.6019854},"labels":[],"label_agreement":null},{"id":"W2089649917","doi":"10.1007/s12561-012-9063-8","title":"Analysis for temporal gene expressions under multiple biological conditions","year":2012,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Gene expression and cancer classification","field":"Biochemistry, Genetics and Molecular Biology","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary; University of Regina","funders":"","keywords":"Mahalanobis distance; Covariance; Data set; Gene expression; Gene; Expression (computer science); Computational biology; Set (abstract data type); Computer science; Biology; Mathematics; Statistics; Genetics; Artificial intelligence","score_opus":0.06634555452040823,"score_gpt":0.3598802756948943,"score_spread":0.29353472117448604,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2089649917","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":[0.57682294,0.00025565742,0.4213159,0.000111091045,0.00031342648,0.00018643244,0.0008334592,0.000008958486,0.0001521335],"genre_scores_gemma":[0.9734172,0.00007575461,0.025270019,0.00015340782,0.0000933082,0.00009726885,0.0007065927,0.000004149443,0.00018230052],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.9991852,0.000050206814,0.0001742643,0.0002482188,0.00010085726,0.00024124538],"domain_scores_gemma":[0.99955124,0.000058714446,0.00007468264,0.00017170017,0.0000509443,0.0000927351],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00022563835,0.00008359869,0.00010034491,0.000118653035,0.0001448354,0.000019638721,0.00015604724,0.00008166369,0.00004617142],"category_scores_gemma":[0.00019755153,0.000065223394,0.000050252205,0.00033920287,0.00018903449,0.0000062004856,0.00004441536,0.000035592206,0.0000027989747],"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":[0.000013261218,0.000113406226,0.36087984,0.0000025271238,0.000022090231,1.0387314e-7,0.00005220771,0.00020563132,0.6337678,0.002664796,0.0020662278,0.00021209146],"study_design_scores_gemma":[0.00050681795,0.00015237201,0.85311687,0.0000056605395,0.000047999896,0.0000014075916,0.0010309239,0.0014270525,0.12483377,0.001857902,0.0167252,0.00029402535],"about_ca_topic_score_codex":0.000014657599,"about_ca_topic_score_gemma":0.00008296035,"teacher_disagreement_score":0.508934,"about_ca_system_score_codex":0.00001199737,"about_ca_system_score_gemma":0.000047729347,"threshold_uncertainty_score":0.26597327},"labels":[],"label_agreement":null},{"id":"W2090437681","doi":"10.1007/s12561-010-9030-1","title":"Sequential Testing with Recurrent Events over Multiple Treatment Periods","year":2010,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":6,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Canadian Institutes of Health Research; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; GlaxoSmithKline","keywords":"Statistics; Medicine; Clinical trial; Null hypothesis; Biostatistics; Baseline (sea); Event (particle physics); Early stopping; Accrual; Statistical hypothesis testing; Econometrics; Computer science; Mathematics; Internal medicine; Epidemiology; Machine learning","score_opus":0.5242442790962877,"score_gpt":0.5618258742737673,"score_spread":0.03758159517747961,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2090437681","genre_codex":"empirical","genre_gemma":"methods","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":[0.81366986,0.000012373762,0.17743987,0.00013796063,0.004397558,0.0010526766,0.0013326285,0.00012349652,0.0018335958],"genre_scores_gemma":[0.34421828,0.0000060098328,0.6555188,0.000019488503,0.0001184452,0.000036870042,0.0000021670405,0.000013801101,0.00006616115],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9974546,0.0002838445,0.0006724377,0.0005425907,0.00058687263,0.0004596827],"domain_scores_gemma":[0.9743374,0.024753148,0.00027063585,0.00035114752,0.00012058902,0.00016707764],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0016799453,0.00024098717,0.00043061146,0.00010795521,0.00017364103,0.00007839202,0.0003386062,0.000103799415,0.00027264902],"category_scores_gemma":[0.062236737,0.00016249648,0.000037434682,0.0004844048,0.00064928597,0.00007861482,0.00007966298,0.00031277657,0.000012881557],"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":[0.00038083087,0.00229612,0.40599525,0.00016607104,0.00007853565,0.00019166748,0.001348557,0.000018652825,0.012488067,0.35887423,0.0007499295,0.21741208],"study_design_scores_gemma":[0.0019862896,0.0014538189,0.062233746,0.00012598051,0.00006561703,0.000014992606,0.00014581761,0.008080654,0.0012760612,0.92365193,0.0004924823,0.00047258445],"about_ca_topic_score_codex":0.00016559706,"about_ca_topic_score_gemma":0.0017452718,"teacher_disagreement_score":0.56477773,"about_ca_system_score_codex":0.00008295154,"about_ca_system_score_gemma":0.00021690872,"threshold_uncertainty_score":0.94566244},"labels":[],"label_agreement":null},{"id":"W2092300754","doi":"10.1007/s12561-010-9021-2","title":"Longitudinal Data Analysis with Event Time as a Covariate","year":2010,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"Natural Sciences and Engineering Research Council of Canada; National Institutes of Health; National Science Foundation","keywords":"Censoring (clinical trials); Covariate; Estimator; Statistics; Econometrics; Nonparametric statistics; Conditional probability distribution; Event (particle physics); Mathematics; Likelihood function; Regression analysis; Estimation theory","score_opus":0.13633690927900644,"score_gpt":0.4425373315305741,"score_spread":0.3062004222515677,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2092300754","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":[0.04904083,0.000005705561,0.9460362,0.00016125628,0.00019442277,0.00015260531,0.0013485212,0.000030339228,0.0030301164],"genre_scores_gemma":[0.29173315,0.0000042750717,0.7079181,0.000030973682,0.000022627022,0.000005405364,0.000027948947,0.0000057866414,0.0002517609],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9982031,0.000093675226,0.00034138787,0.000535947,0.0005092646,0.0003166445],"domain_scores_gemma":[0.9969025,0.001985079,0.0001513147,0.0007401192,0.0001024889,0.000118450036],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0015298079,0.00014781026,0.00031613087,0.0001992924,0.00011454048,0.0001175218,0.000788059,0.000051344876,0.0012887821],"category_scores_gemma":[0.0049495744,0.00010180199,0.000018539342,0.0011950473,0.000533615,0.00010937191,0.00019125934,0.00024284697,0.00005634657],"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":[0.000020970481,0.00014661298,0.018115446,0.00001976669,0.0000687029,0.000051395607,0.00014478114,0.0000071006216,0.00045412153,0.97526914,0.00068350293,0.0050184787],"study_design_scores_gemma":[0.00023376466,0.00022525311,0.062452737,0.000023243192,0.00029784828,0.000013522415,0.00007750271,0.08279701,0.00012231733,0.85314244,0.00031227036,0.00030210428],"about_ca_topic_score_codex":0.00028899263,"about_ca_topic_score_gemma":0.002067154,"teacher_disagreement_score":0.2426923,"about_ca_system_score_codex":0.000013231155,"about_ca_system_score_gemma":0.00016470684,"threshold_uncertainty_score":0.9996242},"labels":[],"label_agreement":null},{"id":"W2095395446","doi":"10.1007/s12561-013-9087-8","title":"Statistical Issues in Modeling Chronic Disease in Cohort Studies","year":2013,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":41,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Biostatistics; Econometrics; Disease; Cohort; Medicine; Observational study; Cohort study; Markov model; Statistics; Computer science; Markov chain; Epidemiology; Machine learning; Mathematics; Internal medicine","score_opus":0.14914579272641043,"score_gpt":0.46776663486276215,"score_spread":0.3186208421363517,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2095395446","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":[0.44913453,0.0014602226,0.5460121,0.00060804497,0.00053017266,0.0010802804,0.00045116295,0.00004988597,0.00067361287],"genre_scores_gemma":[0.5998736,0.00058945024,0.39929715,0.000039193066,0.000026987806,0.000109428154,0.0000043050604,0.000009526492,0.00005037151],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9974231,0.00026008624,0.0007542912,0.00050647446,0.00048107933,0.0005749445],"domain_scores_gemma":[0.9964285,0.0029861908,0.00008148566,0.00023657405,0.00011643385,0.00015082552],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0011942452,0.00021203031,0.00046638,0.00027231188,0.000064300715,0.00008650825,0.00031112751,0.000053982018,0.00028478686],"category_scores_gemma":[0.011579184,0.00017286064,0.000015693473,0.0005619731,0.0005705081,0.00016569812,0.00012462385,0.00025977963,0.00003442197],"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":[0.000010275523,0.00017070558,0.13706264,0.00021595768,0.00000726725,0.00007550701,0.00070469687,0.000491156,0.00004124815,0.85223657,0.00056637573,0.008417616],"study_design_scores_gemma":[0.00016650037,0.000054871743,0.06359433,0.00015108916,0.000006508341,3.575982e-7,0.00046985946,0.28523123,0.0000069908374,0.6501409,0.000009461048,0.00016791937],"about_ca_topic_score_codex":0.0014140462,"about_ca_topic_score_gemma":0.002160665,"teacher_disagreement_score":0.28474006,"about_ca_system_score_codex":0.00022826155,"about_ca_system_score_gemma":0.00019225544,"threshold_uncertainty_score":0.9967467},"labels":[],"label_agreement":null},{"id":"W2125245433","doi":"10.1007/s12561-012-9074-5","title":"On Method of Moments Estimation in Linear Mixed Effects Models with Measurement Error on Covariates and Response with Application to a Longitudinal Study of Gene-Environment Interaction","year":2012,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Lunenfeld-Tanenbaum Research Institute; Mount Sinai Hospital; University of Manitoba; Memorial University of Newfoundland","funders":"","keywords":"Covariate; Observational error; Estimator; Errors-in-variables models; Identifiability; Instrumental variable; Statistics; Random effects model; Mathematics; Moment (physics); Econometrics; Mixed model; Computer science","score_opus":0.10426120790213093,"score_gpt":0.41350311985067895,"score_spread":0.309241911948548,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2125245433","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":[0.37476066,0.000001801045,0.6246591,0.000009753082,0.000024048186,0.0005141651,0.000021225784,0.0000025277066,0.0000067290207],"genre_scores_gemma":[0.5045879,6.341871e-7,0.49534565,0.000003550438,0.0000014465662,0.00005622324,5.753517e-7,0.00000352941,5.117134e-7],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.99827373,0.0004163125,0.00032077538,0.00024684297,0.0005866504,0.00015566869],"domain_scores_gemma":[0.99747014,0.002044162,0.00020855208,0.00016052654,0.000057623274,0.00005899498],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002066043,0.00012439533,0.0002470454,0.00021794777,0.00003344505,0.0000099931485,0.00008116928,0.000025417325,0.0000022518993],"category_scores_gemma":[0.00095480995,0.00008572514,0.00000514683,0.00026720695,0.00007932086,0.000082133556,0.000021232472,0.00007676187,5.6586697e-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":[0.016145607,0.0113261435,0.07219422,0.0007607358,0.000112541784,0.000014493673,0.022765124,0.04150771,0.015263687,0.745932,0.000017071805,0.07396067],"study_design_scores_gemma":[0.002093263,0.012137159,0.5618103,0.0006536437,0.00009514028,0.000004458506,0.002289072,0.21482664,0.014270561,0.19148754,5.5656574e-7,0.00033165852],"about_ca_topic_score_codex":0.00009729146,"about_ca_topic_score_gemma":0.00008760337,"teacher_disagreement_score":0.5544445,"about_ca_system_score_codex":0.00011204679,"about_ca_system_score_gemma":0.000024905392,"threshold_uncertainty_score":0.34957698},"labels":[],"label_agreement":null},{"id":"W2213544608","doi":"10.1007/s12561-015-9134-8","title":"Gene-Environment Independence in Case–Control Studies: Issues of Parameterization and Bayesian Inference","year":2015,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Bayesian probability; Inference; Estimator; Independence (probability theory); Computer science; Bayesian inference; Nuisance parameter; Econometrics; Mathematics; Statistics; Artificial intelligence","score_opus":0.03639241948452713,"score_gpt":0.34321750587347843,"score_spread":0.3068250863889513,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2213544608","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":[0.95561635,0.0014283406,0.042529304,0.00011383337,0.00007808151,0.000112377136,0.000091892485,0.000001274325,0.000028565568],"genre_scores_gemma":[0.9608712,0.0012518396,0.03777129,0.000048614856,0.00001266236,0.000010867963,0.000008610041,0.0000024894186,0.000022386277],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99912876,0.00013313054,0.00025989674,0.00022695892,0.000107477295,0.00014374853],"domain_scores_gemma":[0.99960136,0.000092829425,0.00011032389,0.0001057688,0.000048596954,0.000041107356],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007419383,0.00007315226,0.00015300601,0.000058603495,0.000025644922,0.0000067780607,0.00008371882,0.00007240181,0.0000024109704],"category_scores_gemma":[0.0012382453,0.00006516564,0.0000070046513,0.000093327566,0.00027913885,0.0000043237997,0.00006007726,0.000041751,4.8935294e-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":[0.000014232482,0.00005046983,0.9794019,0.00001761893,0.000015438749,0.000024891497,0.0008981056,0.0030540538,0.012232472,0.00055049034,0.00008107977,0.0036592567],"study_design_scores_gemma":[0.0024564567,0.0022938892,0.88639736,0.000069744245,0.00004921604,0.00012535417,0.012694624,0.039483674,0.01250489,0.042782575,0.00041331287,0.0007289328],"about_ca_topic_score_codex":0.00019630036,"about_ca_topic_score_gemma":0.00070049166,"teacher_disagreement_score":0.093004555,"about_ca_system_score_codex":0.000018596436,"about_ca_system_score_gemma":0.000059407972,"threshold_uncertainty_score":0.26573774},"labels":[],"label_agreement":null},{"id":"W2316257857","doi":"10.1007/s12561-016-9143-2","title":"Power and Effective Study Size in Heritability Studies","year":2016,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":15,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"National Institute of General Medical Sciences; National Institutes of Health; Pacific Institute for the Mathematical Sciences","keywords":"Heritability; Sample size determination; Statistics; Mathematics; Confidence interval; Eigenvalues and eigenvectors; Kinship; Biology; Genetics","score_opus":0.013560786674544127,"score_gpt":0.3355125300005352,"score_spread":0.3219517433259911,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2316257857","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":[0.9984344,0.00031699837,0.0004875198,0.0002155198,0.00013630062,0.00024351587,0.00004289195,0.0000021858002,0.00012061625],"genre_scores_gemma":[0.9964273,0.00026529367,0.0031286182,0.00005310429,0.00001225127,0.00003847129,6.810193e-7,0.0000027311692,0.00007156298],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99902123,0.00021777487,0.00019008324,0.0003184706,0.00006995798,0.00018247844],"domain_scores_gemma":[0.9992707,0.00048680874,0.00005019588,0.000116486626,0.00004901147,0.000026832859],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009445975,0.00007861327,0.00015032396,0.00003103274,0.000044120196,0.000006309356,0.00008650964,0.000045339053,0.000006828307],"category_scores_gemma":[0.0038691736,0.00005064062,0.000009339247,0.00010665335,0.00035152602,0.000002722793,0.00009973519,0.000031812873,0.0000012890845],"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":[0.000011407181,0.00008310471,0.9893693,0.0000029905748,0.000008158628,0.0000018969793,0.00038611246,0.0000016425514,0.0069531156,0.0001779862,0.000112413836,0.002891899],"study_design_scores_gemma":[0.000377808,0.00052408356,0.99141777,0.0000066256034,0.0000020555763,7.216574e-7,0.002309621,0.0000051975207,0.00023995788,0.0049131634,0.00012343166,0.00007959639],"about_ca_topic_score_codex":0.00004350372,"about_ca_topic_score_gemma":0.001513573,"teacher_disagreement_score":0.0067131575,"about_ca_system_score_codex":0.000024778612,"about_ca_system_score_gemma":0.000028732784,"threshold_uncertainty_score":0.46320418},"labels":[],"label_agreement":null},{"id":"W2559903138","doi":"10.1007/s12561-016-9185-5","title":"Classification of Large-Scale Remote Sensing Images for Automatic Identification of Health Hazards","year":2016,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Remote Sensing in Agriculture","field":"Environmental Science","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":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Remote sensing; Computer science; Scale (ratio); Identification (biology); Satellite; Random forest; Data mining; Land cover; Hazard; Artificial intelligence; Geography; Cartography; Engineering","score_opus":0.015907703578447594,"score_gpt":0.3006333137649468,"score_spread":0.28472561018649917,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2559903138","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":[0.27856538,0.000022395876,0.7193213,0.00095768523,0.00024760934,0.00033367568,0.0002873346,0.000017634315,0.00024699405],"genre_scores_gemma":[0.7612543,0.000026219821,0.2385615,0.000017560185,0.000005566205,2.0281223e-7,0.0000063772504,0.000003895763,0.00012438318],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.99847704,0.00009692165,0.00058009,0.0002673825,0.00036682305,0.00021171707],"domain_scores_gemma":[0.99893236,0.00023076622,0.00055854896,0.00020466778,0.000045348457,0.00002829444],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001357478,0.0000808913,0.0001748197,0.000060722392,0.00008927187,0.000014168613,0.00017923943,0.000039677587,0.000019347057],"category_scores_gemma":[0.0005071744,0.000052223353,0.00002387897,0.00037117917,0.00050873414,0.00011875752,0.00004061851,0.000030359013,0.0000071380855],"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":[0.000003223759,0.000037197227,0.0027148216,0.00006457892,0.0000017347784,1.932383e-7,0.0006701322,0.000038136786,0.55804914,0.00045901694,0.0016638768,0.43629792],"study_design_scores_gemma":[0.0003145911,0.000115667804,0.75462973,0.00020016235,0.000008231522,0.0000029356277,0.0008583421,0.15967225,0.064262725,0.019392272,0.0003771843,0.00016589017],"about_ca_topic_score_codex":0.0001643487,"about_ca_topic_score_gemma":0.00029245962,"teacher_disagreement_score":0.7519149,"about_ca_system_score_codex":0.00012770513,"about_ca_system_score_gemma":0.000041055824,"threshold_uncertainty_score":0.21296066},"labels":[],"label_agreement":null},{"id":"W2735500398","doi":"10.1007/s12561-019-09234-6","title":"Bayesian Sensitivity Analysis for Non-ignorable Missing Data in Longitudinal Studies","year":2019,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":false,"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":"Missing data; Computer science; Bayesian probability; Sensitivity (control systems); Data mining; Econometrics; Statistics; Outcome (game theory); Contrast (vision); Machine learning; Artificial intelligence; Mathematics","score_opus":0.19193641977606773,"score_gpt":0.4641900731555287,"score_spread":0.27225365337946095,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2735500398","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":[0.008141312,0.000076166005,0.9896355,0.00012477013,0.0002590464,0.00034109506,0.0010697132,0.0000132905225,0.00033905148],"genre_scores_gemma":[0.3833655,0.00003512649,0.6164962,0.000022039989,0.000016316571,0.0000073773876,0.000017724331,0.0000057132293,0.000033973283],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99783456,0.00019135409,0.0005013563,0.00068393664,0.0003469098,0.000441903],"domain_scores_gemma":[0.9912539,0.007805603,0.00017197279,0.0005757217,0.00012384311,0.00006895545],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003382792,0.00017751502,0.00061212765,0.00035932258,0.000108578344,0.00010146304,0.00038035458,0.00005810308,0.00004894517],"category_scores_gemma":[0.007346514,0.00014722138,0.000034324417,0.0013397816,0.00031666906,0.00024227769,0.00019552019,0.00013926413,0.0000044341273],"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":[0.00005076597,0.00027400322,0.5721057,0.00077040115,0.00015231354,0.00008232665,0.00096901786,0.00006978063,0.00033649645,0.3785337,0.0006812335,0.045974284],"study_design_scores_gemma":[0.00028148896,0.00008798437,0.056577828,0.000120740755,0.00011573524,0.0000019754636,0.00076640083,0.20324884,0.00009395701,0.7383974,0.000050151946,0.0002575031],"about_ca_topic_score_codex":0.00015689751,"about_ca_topic_score_gemma":0.0037667463,"teacher_disagreement_score":0.51552784,"about_ca_system_score_codex":0.000068208865,"about_ca_system_score_gemma":0.00011331155,"threshold_uncertainty_score":0.87949944},"labels":[],"label_agreement":null},{"id":"W2736760148","doi":"10.1007/s12561-017-9201-4","title":"Second-Order Estimating Equations for Clustered Current Status Data from Family Studies Using Response-Dependent Sampling","year":2017,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Systemic Lupus Erythematosus Research","field":"Medicine","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 Waterloo","funders":"Medical Research Council; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Proband; Disease; Family history; Econometrics; Sampling bias; Statistics; Medicine; Kinship; Sample size determination; Demography; Mathematics; Genetics; Biology; Internal medicine; Mutation","score_opus":0.5146520000595493,"score_gpt":0.5441651331125676,"score_spread":0.029513133053018392,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2736760148","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":[0.47581044,0.002047654,0.5141184,0.00019551707,0.002040073,0.0009083267,0.004805432,0.000026582893,0.00004760247],"genre_scores_gemma":[0.44215536,0.00014825333,0.5571775,0.000031953103,0.00018257178,0.0000396053,0.00013917292,0.000020963585,0.00010463196],"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9972951,0.0001408438,0.00059981697,0.00066610845,0.0007413866,0.0005567662],"domain_scores_gemma":[0.99386066,0.0039792256,0.0003826378,0.0011977864,0.0004410388,0.00013863972],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0025788113,0.00017512622,0.00041484373,0.00022797089,0.0008815971,0.00029373547,0.0008044372,0.00005222981,0.000026985213],"category_scores_gemma":[0.029408421,0.00014535437,0.000022284634,0.00016517554,0.0005738914,0.00031016473,0.0007207929,0.0001961105,0.000009004261],"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":[0.004465502,0.0013804166,0.6590308,0.008148971,0.0009282108,0.00024593313,0.048797738,0.0029041145,0.12371839,0.007976436,0.009446514,0.13295698],"study_design_scores_gemma":[0.0028193076,0.00020472343,0.017572366,0.0022035334,0.00010056025,0.000043839103,0.014546131,0.9517998,0.00019397143,0.00939711,0.00075233413,0.0003663186],"about_ca_topic_score_codex":0.00035671782,"about_ca_topic_score_gemma":0.0011860429,"teacher_disagreement_score":0.9488957,"about_ca_system_score_codex":0.00028266947,"about_ca_system_score_gemma":0.00097075163,"threshold_uncertainty_score":0.9787673},"labels":[],"label_agreement":null},{"id":"W2792491629","doi":"10.1007/s12561-018-9213-8","title":"Estimation of Causal Effect Measures in the Presence of Measurement Error in Confounders","year":2018,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Confounding; Estimator; Statistics; Context (archaeology); Odds ratio; Econometrics; Biostatistics; Estimation; Causal inference; Mathematics; Medicine; Epidemiology; Engineering; Geography; Internal medicine","score_opus":0.21513680053283413,"score_gpt":0.4531936732303652,"score_spread":0.23805687269753106,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2792491629","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":[0.5394073,0.000058826852,0.45754632,0.000108449305,0.00018538976,0.0010305817,0.000049756323,0.00002785628,0.0015854967],"genre_scores_gemma":[0.9233811,0.000008428652,0.07655532,0.000011847717,0.0000055560954,0.000030262789,8.0262015e-7,0.000004118643,0.000002597038],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99798477,0.00029722595,0.00050654355,0.0001832596,0.000824117,0.00020408344],"domain_scores_gemma":[0.99787056,0.0014902856,0.00023919788,0.00022715436,0.00015590423,0.000016918015],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0039351997,0.00010932136,0.00023893436,0.00021905208,0.000029405934,0.000014310413,0.00040269035,0.000047193455,0.00000914722],"category_scores_gemma":[0.0072292415,0.00007499477,0.000013343803,0.0006100002,0.001043704,0.00012734243,0.000037348436,0.0001193779,7.261515e-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":[0.0001340309,0.00050946395,0.042947646,0.0006546474,0.000011290938,0.000015182778,0.016798014,0.0006103955,0.025347324,0.88736016,0.00065016944,0.02496165],"study_design_scores_gemma":[0.0003682535,0.0009021262,0.024523905,0.0005288508,0.000010485511,0.0000021048418,0.0016522717,0.017761545,0.0603406,0.8937307,0.000008562374,0.00017059612],"about_ca_topic_score_codex":0.00047762124,"about_ca_topic_score_gemma":0.007380168,"teacher_disagreement_score":0.38397375,"about_ca_system_score_codex":0.00008667788,"about_ca_system_score_gemma":0.00011959341,"threshold_uncertainty_score":0.86546},"labels":[],"label_agreement":null},{"id":"W2955682081","doi":"10.1007/s12561-019-09250-6","title":"A Review of Statistical Analyses on Physical Activity Data Collected from Accelerometers","year":2019,"lang":"en","type":"review","venue":"Statistics in Biosciences","topic":"Physical Activity and Health","field":"Medicine","cited_by":14,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"National Cancer Institute; Natural Sciences and Engineering Research Council of Canada","keywords":"Accelerometer; Raw data; Physical activity; Computer science; Wearable computer; Variety (cybernetics); Energy expenditure; Biostatistics; Data mining; Data science; Physical medicine and rehabilitation; Artificial intelligence; Medicine; Embedded system; Epidemiology","score_opus":0.5743574225043262,"score_gpt":0.5813326418954523,"score_spread":0.0069752193911261084,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2955682081","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":[0.00017235159,0.9604142,0.0012976894,0.00006840827,0.00023771946,0.0013219041,0.03594047,0.000018619596,0.00052866916],"genre_scores_gemma":[0.00040387033,0.9924536,0.005058442,0.00015713852,0.00009321951,0.000019955201,0.0017648127,0.000019288125,0.000029671222],"study_design_codex":"design_other","study_design_gemma":"systematic_review","domain_scores_codex":[0.99692905,0.00042360634,0.00057675276,0.0009171016,0.0007926703,0.0003608219],"domain_scores_gemma":[0.993582,0.004526997,0.0005884821,0.0010437166,0.00007963779,0.00017914992],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.00039755175,0.0003630788,0.003739022,0.0002660628,0.00005264626,0.000023492981,0.00072878937,0.0001245141,0.00010548271],"category_scores_gemma":[0.0022975502,0.00024752907,0.00013995622,0.0014926153,0.0005129748,0.000114429226,0.00026015757,0.00068473036,0.000051406645],"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":[0.00004424071,0.00210833,0.00000696013,0.13996354,0.00012330989,0.000023827744,0.000024470992,8.544443e-8,0.0000134300335,0.0009546686,0.006056733,0.8506804],"study_design_scores_gemma":[0.0010988144,0.004222161,0.0029650705,0.5286712,0.006635377,0.000012423986,0.000033919077,0.0029880852,0.000036882586,0.002555852,0.4494661,0.0013140615],"about_ca_topic_score_codex":0.0005466217,"about_ca_topic_score_gemma":0.00008005293,"teacher_disagreement_score":0.84936637,"about_ca_system_score_codex":0.00016326102,"about_ca_system_score_gemma":0.0017188016,"threshold_uncertainty_score":0.9999977},"labels":[],"label_agreement":null},{"id":"W3010247233","doi":"10.1007/s12561-020-09270-7","title":"A Mixture Model for Bivariate Interval-Censored Failure Times with Dependent Susceptibility","year":2020,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Canadian Institutes of Health Research; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Bivariate analysis; Copula (linguistics); Statistics; Econometrics; Mathematics; Inference; Likelihood function; Accelerated failure time model; Confidence interval; Population; Survival analysis; Maximum likelihood; Computer science; Medicine","score_opus":0.06900862316380572,"score_gpt":0.36422775930383877,"score_spread":0.29521913614003303,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3010247233","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":[0.0019863255,0.000008501253,0.9943991,0.0012119252,0.00007033887,0.00047547286,0.0015908825,0.0000482181,0.0002092377],"genre_scores_gemma":[0.27884743,0.0000035357978,0.7208283,0.00019702368,0.000025785977,0.00002969236,0.0000070726683,0.000011777937,0.000049413316],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9981857,0.00010868662,0.00040508484,0.0005473002,0.00037172294,0.00038152726],"domain_scores_gemma":[0.99785286,0.0014584385,0.00015043611,0.00021123733,0.00015888845,0.000168108],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005588061,0.00020847132,0.00036451526,0.000057443624,0.000109616456,0.000107283005,0.0003841273,0.0000794454,0.00011062254],"category_scores_gemma":[0.0036973718,0.000144515,0.0000343015,0.00032093786,0.00037349266,0.000096333264,0.00007672899,0.00018390636,0.0000045762586],"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":[0.00012949627,0.000105660125,0.0010252912,0.00025938282,0.000012763188,0.000013146058,0.0019398648,0.000050985065,0.0006479207,0.98793656,0.0024173714,0.005461572],"study_design_scores_gemma":[0.00037676707,0.00031440775,0.00026563404,0.00004626369,0.000024156308,0.000001925134,0.0003548345,0.3020891,0.00029421056,0.69598854,0.00004460911,0.00019958202],"about_ca_topic_score_codex":0.000021501102,"about_ca_topic_score_gemma":0.0005699515,"teacher_disagreement_score":0.3020381,"about_ca_system_score_codex":0.000036346853,"about_ca_system_score_gemma":0.00015621909,"threshold_uncertainty_score":0.58931506},"labels":[],"label_agreement":null},{"id":"W312153144","doi":"10.1007/s12561-016-9150-3","title":"Incorporating High-Dimensional Exposure Modelling into Studies of Air Pollution and Health","year":2016,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Air Quality and Health Impacts","field":"Environmental Science","cited_by":9,"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","funders":"University of Bath","keywords":"Air pollution; Biostatistics; Environmental health; Environmental science; Public health; Pollution; Environmental planning; Medicine; Chemistry; Biology","score_opus":0.09532752006570737,"score_gpt":0.36708333810790944,"score_spread":0.27175581804220206,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W312153144","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":[0.8715912,0.00039017954,0.11333849,0.014174119,0.0002035664,0.00016335947,0.000080103244,0.00001317994,0.000045772867],"genre_scores_gemma":[0.82630783,0.00017514604,0.17272271,0.0007612366,0.000009487645,0.0000018925356,8.3351824e-7,0.0000020159396,0.000018871799],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9988979,0.00007975097,0.00032856068,0.00020665486,0.0002864447,0.00020070784],"domain_scores_gemma":[0.99946946,0.00015959874,0.00019799924,0.000073781164,0.000009847542,0.000089329675],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010801571,0.00007151013,0.00015291286,0.00004733569,0.0002028531,0.000004664618,0.00008190826,0.000026185868,0.000016619586],"category_scores_gemma":[0.00012978588,0.000046284826,0.000005533841,0.00019946313,0.00091290026,0.00017783632,0.00010146223,0.000044003096,0.0000036392803],"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":[0.00008714937,0.00029008443,0.31437466,0.0005329837,0.000016194404,0.000010053067,0.019382235,0.054552115,0.008407585,0.2564021,0.0045486973,0.34139615],"study_design_scores_gemma":[0.00072839967,0.0012075081,0.35975897,0.0004652274,0.0000047829967,0.000004164832,0.0028079147,0.034645762,0.0013359636,0.5985913,0.0001292577,0.00032074965],"about_ca_topic_score_codex":0.0009822474,"about_ca_topic_score_gemma":0.0010826788,"teacher_disagreement_score":0.3421892,"about_ca_system_score_codex":0.00010432105,"about_ca_system_score_gemma":0.00005950615,"threshold_uncertainty_score":0.3363619},"labels":[],"label_agreement":null},{"id":"W3179031429","doi":"10.1007/s12561-023-09398-2","title":"Nonnegative Matrix Factorization with Group and Basis Restrictions","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University; University of British Columbia, Okanagan Campus; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Foundation for Innovation","keywords":"Non-negative matrix factorization; Matrix (chemical analysis); Matrix decomposition; Row; Mathematics; Nonnegative matrix; Group (periodic table); Row and column spaces; Rank (graph theory); Basis (linear algebra); Set (abstract data type); Dimensionality reduction; Extension (predicate logic); Curse of dimensionality; Combinatorics; Algorithm; Symmetric matrix; Computer science; Artificial intelligence; Eigenvalues and eigenvectors; Statistics","score_opus":0.018939194728306217,"score_gpt":0.283996933655035,"score_spread":0.26505773892672874,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3179031429","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":[0.07507785,0.000012410295,0.92350024,0.00041634092,0.00029048877,0.00011529787,0.00007536113,0.00010839929,0.00040363998],"genre_scores_gemma":[0.8642113,0.00017162366,0.13541478,0.000035976154,0.000012927829,0.000015575468,0.00001451011,0.0000026948521,0.000120569945],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9992444,0.000034686338,0.00010405403,0.0002437443,0.00022124604,0.00015185999],"domain_scores_gemma":[0.99957305,0.00020480464,0.000046085945,0.00009505989,0.00003686425,0.000044155113],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00015609001,0.00006102559,0.000060769748,0.00023646049,0.00017841694,0.00013848541,0.00017447148,0.000024019084,0.0000057490465],"category_scores_gemma":[0.00008653798,0.000045749068,0.000004486626,0.0015014559,0.00011355868,0.0003851317,0.00006745911,0.00005301814,0.000017647762],"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":[0.000025689711,0.00018023532,0.13447316,0.00008326213,0.000014900524,0.00010907777,0.009640631,0.0007576696,0.009022361,0.7272462,0.010265626,0.10818118],"study_design_scores_gemma":[0.0008004884,0.00068403,0.5602642,0.0001718874,0.000009620915,0.00001380457,0.0033435826,0.27226558,0.0029049066,0.15733735,0.001627219,0.0005772757],"about_ca_topic_score_codex":0.00009622284,"about_ca_topic_score_gemma":0.00019488203,"teacher_disagreement_score":0.7891335,"about_ca_system_score_codex":0.000013310121,"about_ca_system_score_gemma":0.000032040494,"threshold_uncertainty_score":0.18655929},"labels":[],"label_agreement":null},{"id":"W3200223375","doi":"10.1007/s12561-021-09320-8","title":"Is 14-Days a Sensible Quarantine Length for COVID-19? Examinations of Some Associated Issues with a Case Study of COVID-19 Incubation Times","year":2021,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Canada Excellence Research Chairs, Government of Canada; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Coronavirus disease 2019 (COVID-19); 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Quarantine; Incubation period; Virology; Medicine; Incubation; Statistics; Biology; Mathematics; Outbreak; Internal medicine; Infectious disease (medical specialty); Pathology","score_opus":0.2963591884773393,"score_gpt":0.49294542748967524,"score_spread":0.19658623901233596,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3200223375","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":[0.9010587,0.00022444685,0.091961734,0.0031315952,0.000103691986,0.0012680953,0.0021280427,0.00006571146,0.000057998626],"genre_scores_gemma":[0.9543702,0.0000934226,0.044676337,0.0005535168,0.000015956923,0.000074230404,0.000028315182,0.000011040101,0.0001769669],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9972055,0.0005482056,0.00087171583,0.0005148078,0.00054741226,0.0003123826],"domain_scores_gemma":[0.98289084,0.015419325,0.0006986406,0.00028675236,0.0005721321,0.00013232414],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0029511224,0.00020676135,0.00070518604,0.00023635756,0.0003495491,0.000029055034,0.00020625992,0.000086750144,0.000063230924],"category_scores_gemma":[0.07873424,0.00015850065,0.00004645775,0.0011619342,0.0005992842,0.00011471375,0.00013851884,0.000111101406,6.6441726e-7],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002621027,0.010105446,0.38576406,0.0042616753,0.0006608385,0.001481821,0.12453134,0.0026553024,0.00095235655,0.4387701,0.029447697,0.0011072645],"study_design_scores_gemma":[0.0057555307,0.0044581974,0.020105062,0.00016735605,0.00046445083,0.00008974115,0.12875655,0.02202791,0.0013427899,0.815244,0.0007525775,0.0008358415],"about_ca_topic_score_codex":0.0042821695,"about_ca_topic_score_gemma":0.021989638,"teacher_disagreement_score":0.3764739,"about_ca_system_score_codex":0.00020277771,"about_ca_system_score_gemma":0.00066303293,"threshold_uncertainty_score":0.9958565},"labels":[],"label_agreement":null},{"id":"W4210355977","doi":"10.1007/s12561-021-09322-6","title":"Correction to: Is 14-Days a Sensible Quarantine Length for COVID-19? Examinations of Some Associated Issues with a Case Study of COVID-19 Incubation Times","year":2021,"lang":"en","type":"erratum","venue":"Statistics in Biosciences","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Coronavirus disease 2019 (COVID-19); Quarantine; 2019-20 coronavirus outbreak; Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2); Virology; Medicine; Biostatistics; Epidemiology; Internal medicine; Pathology; Infectious disease (medical specialty); Outbreak","score_opus":0.21183987229801318,"score_gpt":0.4736330072229911,"score_spread":0.2617931349249779,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4210355977","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":[0.22135396,0.0030118902,0.66581947,0.0151513405,0.033119228,0.022056667,0.03697483,0.0007685602,0.0017440631],"genre_scores_gemma":[0.8524774,0.0014922871,0.0989628,0.0033435745,0.0005736597,0.0011825449,0.0012656811,0.00015675274,0.04054532],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9953437,0.00084775104,0.0014276479,0.0009528808,0.000964646,0.00046341878],"domain_scores_gemma":[0.9774227,0.01929519,0.0016132953,0.00044527638,0.0010016648,0.00022185151],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.003977059,0.00045068486,0.0014886241,0.0006825062,0.00051024975,0.00005939988,0.00037605694,0.00028840502,0.00007522882],"category_scores_gemma":[0.15066119,0.00036188628,0.00009099759,0.001910318,0.0006034785,0.00011736971,0.00023470563,0.0004148395,0.0000010361055],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00008702151,0.001974741,0.0070000333,0.0020260974,0.00024691172,0.00024084224,0.024980523,0.0006571074,0.000021443055,0.008351854,0.9540926,0.00032082296],"study_design_scores_gemma":[0.01174506,0.029519323,0.015191163,0.0033016107,0.0033236519,0.00031466316,0.3322093,0.08413149,0.00039513086,0.46543235,0.049330212,0.0051060715],"about_ca_topic_score_codex":0.01637209,"about_ca_topic_score_gemma":0.08646329,"teacher_disagreement_score":0.9047624,"about_ca_system_score_codex":0.00065126014,"about_ca_system_score_gemma":0.0017041998,"threshold_uncertainty_score":0.9998833},"labels":[],"label_agreement":null},{"id":"W4210899445","doi":"10.1007/s12561-022-09337-7","title":"Building an External Control Arm for Development of a New Molecular Entity: An Application in a Recurrent Glioblastoma Trial for MDNA55","year":2022,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":8,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Pharma Medica Research (Canada)","funders":"","keywords":"Randomized controlled trial; Observational study; Medicine; Clinical trial; Selection bias; Protocol (science); Propensity score matching; Drug development; Biostatistics; Medical physics; Physical therapy; Surgery; Alternative medicine; Internal medicine; Drug; Epidemiology; Pathology; Psychiatry","score_opus":0.1136866173941476,"score_gpt":0.4534375624280227,"score_spread":0.3397509450338751,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4210899445","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":[0.28379613,0.000011717302,0.7144804,0.0000152526945,0.00009592585,0.0013973218,0.0001772253,0.000023127526,0.0000029298842],"genre_scores_gemma":[0.44415352,0.000001041519,0.55504197,0.000012953217,0.000014528117,0.00075830333,0.00000710194,0.0000071708278,0.0000034143538],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99844885,0.000070072245,0.00055455347,0.00033134248,0.00034846025,0.00024671547],"domain_scores_gemma":[0.9990009,0.00037565423,0.000312529,0.00017217203,0.00007602814,0.00006269614],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012805158,0.00011635263,0.00023908912,0.00023919123,0.00015720612,0.000024719206,0.00038956918,0.00003104364,0.000007883118],"category_scores_gemma":[0.00063685933,0.000110482644,0.000024057026,0.00028958297,0.00007934241,0.00013686034,0.00007996541,0.00009333965,5.727551e-8],"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":[0.0041582547,0.0009937974,0.000110577865,0.000104521016,0.000006063472,0.0000023441103,0.0026410765,0.00019156541,0.03464485,0.82176554,0.000037091464,0.13534433],"study_design_scores_gemma":[0.0059316796,0.0028974183,0.000044143653,0.000034268036,0.000009984362,0.0000010988769,0.00036309622,0.022348281,0.019981742,0.94772404,0.00048915995,0.00017505663],"about_ca_topic_score_codex":0.00003840998,"about_ca_topic_score_gemma":0.000843969,"teacher_disagreement_score":0.16035739,"about_ca_system_score_codex":0.00019011275,"about_ca_system_score_gemma":0.00032083795,"threshold_uncertainty_score":0.45053515},"labels":[],"label_agreement":null},{"id":"W4362591564","doi":"10.1007/s12561-023-09369-7","title":"Evaluation of Designs and Estimation Methods Under Response-Dependent Two-Phase Sampling for Genetic Association Studies","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Optimal Experimental Design Methods","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Statistics; Covariate; Sampling (signal processing); Computer science; Robustness (evolution); Sample size determination; Trait; Econometrics; Mathematics; Data mining; Biology; Genetics","score_opus":0.6738625109802414,"score_gpt":0.6789437062369973,"score_spread":0.005081195256755877,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4362591564","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":[0.2590231,0.00042224623,0.73931974,0.00014237984,0.0004578,0.00048945734,0.000106182,0.000017826096,0.000021278125],"genre_scores_gemma":[0.32236898,0.00003774114,0.6774257,0.000021518315,0.0000105337385,0.00006613656,0.0000027347444,0.0000059153604,0.000060762384],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9918636,0.0037788143,0.00088250934,0.0005458506,0.002639513,0.00028972072],"domain_scores_gemma":[0.96576583,0.032339633,0.00051515555,0.00020473768,0.0011203943,0.000054258628],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.08639104,0.0001258685,0.00031211463,0.00065456965,0.00020591554,0.00015506896,0.0002937795,0.000057268935,0.00001922712],"category_scores_gemma":[0.108783215,0.00010332357,0.000030437819,0.001485112,0.00026551407,0.00023312662,0.000102623344,0.00006044493,0.0000064913734],"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":[0.0002726194,0.00011482835,0.0017866155,0.000022140386,0.000044856206,0.0000016250445,0.005650919,0.08796117,0.20960079,0.00790491,0.00040176063,0.68623775],"study_design_scores_gemma":[0.00083452766,0.0002856819,0.014713103,0.000016214173,0.000044444656,9.788041e-7,0.0057147034,0.5859669,0.014332578,0.3779759,0.00001131411,0.00010364382],"about_ca_topic_score_codex":0.000018997762,"about_ca_topic_score_gemma":0.000041099785,"teacher_disagreement_score":0.6861341,"about_ca_system_score_codex":0.00032710045,"about_ca_system_score_gemma":0.0002632065,"threshold_uncertainty_score":0.9407527},"labels":[],"label_agreement":null},{"id":"W4379206837","doi":"10.1007/s12561-023-09373-x","title":"Variable Selection in Multivariate Functional Linear Regression","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Reproducing kernel Hilbert space; Mathematics; Multivariate statistics; Estimator; Covariate; Functional data analysis; Smoothness; Applied mathematics; Mathematical optimization; Hilbert space; Statistics","score_opus":0.15440883130429617,"score_gpt":0.4307915064557374,"score_spread":0.2763826751514412,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4379206837","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":[0.026217071,0.000006566852,0.9704804,0.000104190956,0.0006709861,0.0001792385,0.00016470246,0.00008978824,0.0020870606],"genre_scores_gemma":[0.087834574,0.000012399836,0.9115051,0.00003263733,0.00004870502,0.000022432625,0.000009338725,0.00000820067,0.0005266352],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99854046,0.00015917455,0.00034251506,0.0003018326,0.0003408164,0.0003152246],"domain_scores_gemma":[0.9973533,0.0023565171,0.00008426739,0.000081992555,0.000071135175,0.000052758875],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013925739,0.00010860737,0.00017483774,0.00028176577,0.0001051796,0.00003293895,0.0001325965,0.000069189555,0.00026798612],"category_scores_gemma":[0.0067930943,0.00008484458,0.000011134196,0.0016756151,0.00013713702,0.00007774353,0.000058156133,0.00018589669,0.000043639237],"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":[0.000015989617,0.00006537896,0.008499884,0.000036130976,0.0000015211325,0.000009531433,0.00017813615,0.00013692837,0.002665343,0.9827986,0.0017370968,0.003855483],"study_design_scores_gemma":[0.00019900591,0.00005266385,0.06292301,0.000072498886,0.0000024164806,0.0000014545313,0.00012916663,0.1524369,0.00023298529,0.78367656,0.00016887893,0.000104486535],"about_ca_topic_score_codex":0.00018728033,"about_ca_topic_score_gemma":0.00016039249,"teacher_disagreement_score":0.19912204,"about_ca_system_score_codex":0.000055616194,"about_ca_system_score_gemma":0.00010903998,"threshold_uncertainty_score":0.8132459},"labels":[],"label_agreement":null},{"id":"W4382795050","doi":"10.1007/s12561-023-09375-9","title":"Statistical Learning of Large-Scale Genetic Data: How to Run a Genome-Wide Association Study of Gene-Expression Data Using the 1000 Genomes Project Data","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","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":"SickKids Foundation; Hospital for Sick Children; Public Health Ontario; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research; University of Toronto","keywords":"Biostatistics; Genetic data; Genome-wide association study; Computational biology; Genome; Scale (ratio); Data mining; Genetic association; Computer science; Data science; Biology; Gene; Genetics; Single-nucleotide polymorphism; Geography; Cartography; Medicine; Population; Public health; Genotype","score_opus":0.07540431863356317,"score_gpt":0.37668960225481,"score_spread":0.30128528362124685,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4382795050","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":[0.86689484,0.00019773327,0.1220556,0.00019842813,0.00018660283,0.0005384893,0.009902251,0.000008183612,0.000017901553],"genre_scores_gemma":[0.88299173,0.00035283595,0.111872226,0.000052422372,0.00008746338,0.000010822799,0.00450223,0.000015293737,0.000114952054],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99712276,0.00059481134,0.0005362573,0.000857441,0.00045633843,0.0004324164],"domain_scores_gemma":[0.99700737,0.0006486902,0.0004514489,0.0016685697,0.00017099922,0.00005291822],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0034604818,0.00014913107,0.0003083495,0.00013993641,0.00022091052,0.000042521424,0.0021428687,0.000102462014,0.0000065602994],"category_scores_gemma":[0.0057429345,0.00011661631,0.0000114492705,0.00073706696,0.00013585872,0.00001924284,0.0026103088,0.00013244731,0.0000019069023],"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":[0.000019497782,0.00018519835,0.9546006,0.000026239743,0.000049607697,0.000003138591,0.0010348773,0.003301532,0.032995943,0.000012917577,0.007207305,0.000563164],"study_design_scores_gemma":[0.0005398709,0.0006128077,0.89711505,0.00001845614,0.00009323376,0.0000013938094,0.0127684465,0.08243642,0.00063972943,0.00009397991,0.00540262,0.00027801134],"about_ca_topic_score_codex":0.00030561967,"about_ca_topic_score_gemma":0.0015178195,"teacher_disagreement_score":0.07913489,"about_ca_system_score_codex":0.00003242838,"about_ca_system_score_gemma":0.000330394,"threshold_uncertainty_score":0.68752444},"labels":[],"label_agreement":null},{"id":"W4385766135","doi":"10.1007/s12561-023-09382-w","title":"Understanding Effective Virus Control Policies for Covid-19 with the Q-learning Method","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"COVID-19 epidemiological studies","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Case fatality rate; Per capita; Coronavirus disease 2019 (COVID-19); Gross domestic product; Environmental health; Medicine; Biostatistics; Population; Public health; Demography; Economic growth; Disease; Economics; Infectious disease (medical specialty)","score_opus":0.470793108057292,"score_gpt":0.5283827705085856,"score_spread":0.0575896624512936,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4385766135","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":[0.0027779636,0.000028866567,0.9883072,0.0072388756,0.00009973632,0.00093786186,0.0002371082,0.00014041926,0.0002320049],"genre_scores_gemma":[0.9253754,0.00006307401,0.07152326,0.0023825758,0.000059276645,0.0004044098,0.000004974473,0.000016635335,0.00017037881],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99802476,0.00057002314,0.00025808087,0.00035692257,0.00031026735,0.00047994882],"domain_scores_gemma":[0.9270152,0.072553575,0.00018142925,0.00012398073,0.000047511712,0.00007831046],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.005122396,0.0001647121,0.00035640414,0.00014320915,0.0008808045,0.00006910378,0.00030172002,0.000050248535,0.000010840307],"category_scores_gemma":[0.041301724,0.00008703439,0.000040027666,0.00092502375,0.00078635296,0.00004677109,0.000098160745,0.00018539105,0.0000043860277],"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":[0.00009232619,0.000026049438,0.0140330745,0.0001602739,0.00004742865,0.000010970861,0.00457343,0.0054275445,0.00023073975,0.9641925,0.010657228,0.00054840575],"study_design_scores_gemma":[0.00072418107,0.00045871697,0.007703585,0.000026919772,0.0000458976,0.0000018759415,0.011021068,0.033385698,0.00003882989,0.94175434,0.00462454,0.00021436132],"about_ca_topic_score_codex":0.00050746667,"about_ca_topic_score_gemma":0.0025045744,"teacher_disagreement_score":0.92259747,"about_ca_system_score_codex":0.00033091233,"about_ca_system_score_gemma":0.000097262884,"threshold_uncertainty_score":0.9667738},"labels":[],"label_agreement":null},{"id":"W4389208232","doi":"10.1007/s12561-023-09407-4","title":"Evaluating Effects of Various Exposures on Mortality Risk of Opioid Use Disorders with Linked Administrative Databases","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Insurance, Mortality, Demography, Risk Management","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Centre for Advancing Health Outcomes; St. Paul's Hospital; Simon Fraser University","funders":"National Institute of Nursing Research; Natural Sciences and Engineering Research Council of Canada","keywords":"Biostatistics; Covariate; Medicine; Hazard; Proportional hazards model; Data mining; Econometrics; Actuarial science; Computer science; Statistics; Public health; Machine learning; Business; Mathematics","score_opus":0.09982524946763711,"score_gpt":0.4270462260048003,"score_spread":0.3272209765371632,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4389208232","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":[0.99500376,0.00003852303,0.0017796621,0.000026737649,0.000443494,0.0008040328,0.00089885714,0.000046943274,0.00095797074],"genre_scores_gemma":[0.9872441,0.00093760213,0.011652088,0.000018259187,0.000022425686,0.000045651046,0.000029679843,0.0000084662815,0.00004169598],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9965315,0.00073036243,0.00046900962,0.000448617,0.0014183476,0.00040217632],"domain_scores_gemma":[0.9962424,0.0026782097,0.0005101118,0.0003382505,0.0001562747,0.00007473324],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002636761,0.00016970055,0.00030683313,0.00032597198,0.00036371077,0.00006191286,0.00043301802,0.00004404436,0.000012661968],"category_scores_gemma":[0.0037228474,0.00013919207,0.000044948127,0.0018160904,0.0022319616,0.00024066387,0.0000794842,0.00014837312,0.0000026495857],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00004077392,0.00022371188,0.9215611,0.00017438474,0.000049475013,0.000014793001,0.007938111,0.00056749745,0.000086785214,0.06513088,0.000097895,0.0041145873],"study_design_scores_gemma":[0.0003282763,0.0007571192,0.9819929,0.0001489266,0.0000621788,4.681407e-8,0.004633553,0.0008213628,0.0003564618,0.010644118,0.000073310584,0.0001817709],"about_ca_topic_score_codex":0.019449705,"about_ca_topic_score_gemma":0.07045671,"teacher_disagreement_score":0.06043177,"about_ca_system_score_codex":0.00003503048,"about_ca_system_score_gemma":0.0003093546,"threshold_uncertainty_score":0.98707986},"labels":[],"label_agreement":null},{"id":"W4389238863","doi":"10.1007/s12561-023-09403-8","title":"Use of Generalized Propensity Scores for Assessing Effects of Multiple Exposures","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"National Institute on Alcohol Abuse and Alcoholism","keywords":"Confounding; Statistics; Conditional independence; Propensity score matching; Econometrics; Causal inference; Bivariate analysis; Joint probability distribution; Estimator; Inverse probability weighting; Causal model; Regression analysis; Marginal structural model; Weighting; Biostatistics; Conditional probability distribution; Mathematics; Medicine; Epidemiology","score_opus":0.32523508943013857,"score_gpt":0.4596776250189145,"score_spread":0.13444253558877595,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4389238863","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":[0.71102214,0.000011151875,0.28812987,0.00000936243,0.00011998795,0.00049651216,0.00013210668,0.00006700016,0.000011898925],"genre_scores_gemma":[0.55129766,0.000023651652,0.44860357,0.000005055305,0.0000052259675,0.000030193223,0.000005044211,0.000006258756,0.000023377164],"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9988723,0.0000751402,0.00038392394,0.00019574208,0.00026143965,0.00021150283],"domain_scores_gemma":[0.99460644,0.0047275065,0.0002754152,0.00018560656,0.0001774061,0.000027650844],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0005358119,0.00010213376,0.00030436183,0.00019554276,0.00005604538,0.00003135736,0.00019425424,0.000045926263,0.0000021326712],"category_scores_gemma":[0.010845004,0.00008157804,0.000027467593,0.0004757378,0.00041026602,0.00022264749,0.00008181602,0.000052023563,3.189408e-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":[0.00004637286,0.00020446336,0.11773747,0.001900183,0.000015242574,0.000014065033,0.0010818309,0.00011548386,0.4529205,0.41593772,0.00208439,0.007942295],"study_design_scores_gemma":[0.00023842482,0.0001781279,0.031693958,0.00026463345,0.000012138314,4.4384237e-7,0.0000938823,0.0064952383,0.45015848,0.51071316,0.000020332283,0.00013118664],"about_ca_topic_score_codex":0.000115474664,"about_ca_topic_score_gemma":0.00024087956,"teacher_disagreement_score":0.16047369,"about_ca_system_score_codex":0.000018671706,"about_ca_system_score_gemma":0.000069059264,"threshold_uncertainty_score":0.99748707},"labels":[],"label_agreement":null},{"id":"W4389266541","doi":"10.1007/s12561-023-09402-9","title":"Causal Mediation Tree Model for Feature Identification on High-Dimensional Mediators","year":2023,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Data Analysis with R","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Princess Margaret Cancer Centre; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Crohn's and Colitis Canada","keywords":"Identification (biology); Feature (linguistics); Mediation; Tree (set theory); Computer science; Data mining; Artificial intelligence; Mathematics; Political science; Biology","score_opus":0.0240054500936361,"score_gpt":0.30508878880325463,"score_spread":0.28108333870961855,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4389266541","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":[0.01944843,0.000009644087,0.97530824,0.0022973008,0.0013129476,0.0002276909,0.0012431348,0.000120014236,0.00003260311],"genre_scores_gemma":[0.8131551,0.000018542918,0.18528439,0.00023924865,0.0000703346,0.000055023545,0.0007115437,0.0000076154474,0.0004581581],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99807733,0.00004489416,0.00027419446,0.00053583196,0.000780434,0.00028731595],"domain_scores_gemma":[0.9987934,0.0004960554,0.00015751392,0.00036391462,0.00011392712,0.000075193486],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009573733,0.00011511431,0.0001413374,0.0004945019,0.00017295948,0.0001765286,0.00084098923,0.000058617763,0.0000026091006],"category_scores_gemma":[0.00075690105,0.00009767823,0.000023715405,0.001548778,0.00011986074,0.00042662525,0.00011226734,0.00008613735,0.000059612747],"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":[0.000013269785,0.00010254087,0.0014356982,0.000023204953,0.000013813889,0.000015982758,0.0011126787,0.0419178,0.0030451247,0.8312533,0.09168811,0.029378487],"study_design_scores_gemma":[0.00015231036,0.00004034451,0.020343741,0.0000054351526,0.0000047180506,3.375295e-7,0.000020224255,0.9130134,0.0005837515,0.06567437,0.00004677042,0.00011459899],"about_ca_topic_score_codex":0.000027388072,"about_ca_topic_score_gemma":0.00053044845,"teacher_disagreement_score":0.8710956,"about_ca_system_score_codex":0.00006304162,"about_ca_system_score_gemma":0.00016100318,"threshold_uncertainty_score":0.3983203},"labels":[],"label_agreement":null},{"id":"W4390704883","doi":"10.1007/s12561-023-09412-7","title":"Functional Linear Partial Quantile Regression with Guaranteed Convergence for Neuroimaging Data Analysis","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":4,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Queen's University; University of Alberta","funders":"","keywords":"Quantile; Quantile regression; Outlier; Functional data analysis; Functional magnetic resonance imaging; Computer science; Functional principal component analysis; Diffusion MRI; Artificial intelligence; Mathematics; Machine learning; Econometrics; Magnetic resonance imaging; Psychology; Medicine","score_opus":0.21639386311973474,"score_gpt":0.45123641122414454,"score_spread":0.2348425481044098,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4390704883","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":[0.0041491305,0.00007299188,0.9923468,0.00016540443,0.0006053535,0.0001692963,0.002327252,0.00004858737,0.00011518066],"genre_scores_gemma":[0.29783112,0.000019747727,0.70186365,0.000045705106,0.00006324038,0.000015953789,0.00006189559,0.000009221143,0.000089480745],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9982665,0.000085578795,0.00034961884,0.0006131572,0.00040743127,0.00027772266],"domain_scores_gemma":[0.9960085,0.0033898482,0.00007982299,0.00037645837,0.00008304052,0.00006231887],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0010550786,0.0001403962,0.0002518336,0.00021183511,0.00013629156,0.00013109755,0.00038218687,0.000032566768,0.00021818742],"category_scores_gemma":[0.0027284913,0.0000920178,0.000028201139,0.0011219158,0.0003728824,0.0001438751,0.00010158578,0.00013192541,0.000004930015],"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":[0.000078425575,0.0000769667,0.006939817,0.00022672623,0.000072144016,0.00004548388,0.00024458667,0.00009969173,0.0004703648,0.9783296,0.004589674,0.008826527],"study_design_scores_gemma":[0.00011357881,0.00010475987,0.0019917209,0.00009174569,0.00017648397,0.0000031136167,0.00014370753,0.8509659,0.00016389006,0.14531836,0.00077821815,0.0001485342],"about_ca_topic_score_codex":0.000058465554,"about_ca_topic_score_gemma":0.00013217212,"teacher_disagreement_score":0.8508662,"about_ca_system_score_codex":0.00001546923,"about_ca_system_score_gemma":0.00015115662,"threshold_uncertainty_score":0.3752377},"labels":[],"label_agreement":null},{"id":"W4399080701","doi":"10.1007/s12561-024-09432-x","title":"Analysis of Recurrent Event Processes with Dynamic Models for Event Counts","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods in Clinical Trials","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Event (particle physics); Computer science; Covariate; Biostatistics; Econometrics; Cluster analysis; Estimation; Parametric statistics; Statistics; Data mining; Machine learning; Mathematics; Medicine; Epidemiology; Engineering","score_opus":0.3809228862421378,"score_gpt":0.5811723911560625,"score_spread":0.20024950491392463,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399080701","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":[0.0010812685,0.00022753075,0.99505293,0.000055214565,0.00049341854,0.00030891475,0.0023433878,0.00002219671,0.00041516198],"genre_scores_gemma":[0.29231542,0.00016255227,0.70730907,0.0000098810415,0.000014213222,0.000045337143,0.000007144079,0.0000087968265,0.0001275707],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9984739,0.00008253604,0.0005504578,0.0003083598,0.000406871,0.00017792202],"domain_scores_gemma":[0.98079026,0.018726801,0.0001332722,0.000149848,0.00015656636,0.000043277465],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0015583835,0.00010302267,0.0003733411,0.00024060076,0.00002892307,0.000036204565,0.00020240387,0.000042220185,0.00008265063],"category_scores_gemma":[0.013494834,0.000070233786,0.00004971767,0.0013065633,0.00026470682,0.000044700948,0.000026869187,0.00007530867,0.000002264588],"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":[0.00006361963,0.00025062662,0.00019890489,0.0019060305,0.00028267762,0.000007878982,0.00042525984,0.001008644,0.000017468185,0.95838,0.0011068707,0.03635205],"study_design_scores_gemma":[0.00007627492,0.00016347707,0.000105402,0.00020563738,0.00031010495,2.9304218e-7,0.00005830832,0.25444186,0.00003241295,0.74442583,0.00009680519,0.00008358145],"about_ca_topic_score_codex":0.000005624267,"about_ca_topic_score_gemma":0.00019687561,"teacher_disagreement_score":0.29123414,"about_ca_system_score_codex":0.000048976606,"about_ca_system_score_gemma":0.0002714375,"threshold_uncertainty_score":0.99481493},"labels":[],"label_agreement":null},{"id":"W4399231422","doi":"10.1007/s12561-024-09435-8","title":"A Comprehensive Performance Comparison Study of Various Statistical Models that Accommodate Challenges of the Gut Microbiome Data","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Gut microbiota and health","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Mitacs","keywords":"Microbiome; Human microbiome; Count data; Computational biology; Biology; Amplicon sequencing; Gut microbiome; Statistical model; Statistics; Computer science; Bioinformatics; Genetics; Poisson distribution; Artificial intelligence; Mathematics; 16S ribosomal RNA","score_opus":0.12466728875127968,"score_gpt":0.37049997287695186,"score_spread":0.24583268412567216,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399231422","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":[0.9896026,0.002825364,0.004098274,0.00012378577,0.0004251656,0.00036518704,0.002424933,0.0000058291675,0.00012881798],"genre_scores_gemma":[0.99428797,0.0011009851,0.0044389297,0.000025441002,0.000017412756,0.0000041513536,0.000098786746,0.000008749033,0.000017591063],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.99863213,0.00012107504,0.00036871585,0.0004435416,0.00020843052,0.00022610577],"domain_scores_gemma":[0.9990771,0.00009203482,0.000118214964,0.0006093333,0.00006784189,0.00003549224],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000320123,0.0001329488,0.00024244825,0.0000788402,0.00007560107,0.000026497035,0.00086069247,0.00006295768,0.0000054787606],"category_scores_gemma":[0.00002944551,0.00009330525,0.000015981519,0.00020851393,0.00046725667,0.000012825849,0.0005747591,0.0001322722,0.000001277889],"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":[0.00028898774,0.0031294806,0.04157319,0.003442934,0.00025900695,0.000023036251,0.013271791,0.0024464983,0.8510612,0.03546152,0.008172381,0.04086993],"study_design_scores_gemma":[0.0032819756,0.0066201044,0.6542313,0.0010319317,0.00026719776,0.00008380783,0.026555449,0.24768734,0.039349116,0.0068736873,0.012498119,0.001519938],"about_ca_topic_score_codex":0.0002778012,"about_ca_topic_score_gemma":0.001166249,"teacher_disagreement_score":0.81171215,"about_ca_system_score_codex":0.000011852468,"about_ca_system_score_gemma":0.00019880457,"threshold_uncertainty_score":0.38048777},"labels":[],"label_agreement":null},{"id":"W4399534806","doi":"10.1007/s12561-024-09436-7","title":"Sieve Estimation of the Additive Hazards Model with Bivariate Current Status Data","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary; University of Alberta","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Copula (linguistics); Estimator; Mathematics; Asymptotic distribution; Bivariate analysis; Applied mathematics; Mathematical optimization; Statistics; Econometrics","score_opus":0.13497020891454667,"score_gpt":0.4406628682488462,"score_spread":0.30569265933429957,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4399534806","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":[0.0032918046,0.00012985198,0.9853024,0.000088150104,0.00042956762,0.00020539772,0.009725333,0.00002290495,0.0008045919],"genre_scores_gemma":[0.3420722,0.00007188521,0.6577598,0.000010489014,0.00001261053,0.000008646885,0.000023842924,0.000007338605,0.000033208828],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9984206,0.00011340811,0.00033720894,0.00035572314,0.00052103004,0.0002520246],"domain_scores_gemma":[0.99726933,0.002036154,0.00011392992,0.00043780505,0.000090054134,0.000052696916],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00075791316,0.00012555937,0.0001819066,0.00008424628,0.00006263571,0.00008713879,0.00052361004,0.00002923997,0.00004378352],"category_scores_gemma":[0.0038230962,0.000069874055,0.000012924273,0.00060974696,0.00058841484,0.00015246394,0.00021847631,0.0001876186,0.0000029156606],"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":[0.000007099044,0.0000475852,0.00009500984,0.00014289525,0.000006500403,0.0000023157788,0.0005710373,0.00045894706,0.000031993874,0.88714194,0.0014667935,0.11002789],"study_design_scores_gemma":[0.000045387365,0.00003814749,0.00079119194,0.00020298836,0.000019037792,7.9423745e-7,0.00007296441,0.5393325,0.000106444044,0.45923766,0.000087399865,0.000065435255],"about_ca_topic_score_codex":0.00006394666,"about_ca_topic_score_gemma":0.0001330571,"teacher_disagreement_score":0.5388736,"about_ca_system_score_codex":0.00004097591,"about_ca_system_score_gemma":0.00055738015,"threshold_uncertainty_score":0.45768794},"labels":[],"label_agreement":null},{"id":"W4400329941","doi":"10.1007/s12561-024-09441-w","title":"Estimation in Multi-State Semi-Markov Models with a Cured Fraction and Masked Causes of Deaths","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Cancer; Medicine; Disease; Inference; Event (particle physics); Estimation; Internal medicine; Computer science; Artificial intelligence","score_opus":0.10724743571891306,"score_gpt":0.410726693157753,"score_spread":0.30347925743883997,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400329941","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":[0.096157536,0.00009367907,0.9030353,0.000030762312,0.00008244844,0.00017196567,0.00023101944,0.000020951255,0.0001763417],"genre_scores_gemma":[0.49411577,0.00006784943,0.5057723,0.000005319493,0.0000017514612,0.000009599017,0.0000018737193,0.0000048077404,0.000020762172],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99886715,0.00009038067,0.0003421733,0.000268954,0.0002522464,0.00017909081],"domain_scores_gemma":[0.99783516,0.0018868434,0.00008346011,0.0000976399,0.00005314821,0.0000437292],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00059875316,0.00011522742,0.00021372727,0.00021202586,0.00003445699,0.000066862944,0.00008321107,0.000040325405,0.000013972508],"category_scores_gemma":[0.0012929233,0.00008522445,0.000008025339,0.00046860767,0.0002981458,0.00018971706,0.000027684677,0.00013025723,7.620163e-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":[0.000057731057,0.00019155772,0.0063231396,0.00086998096,0.0000150811275,0.00006999052,0.003857738,0.0011984124,0.0014504943,0.90567505,0.000088098954,0.08020271],"study_design_scores_gemma":[0.0001173196,0.00007397772,0.0042574853,0.00022347769,0.0000078184175,0.0000029303826,0.00019172812,0.53777874,0.00035452723,0.45691395,0.0000026818298,0.00007535855],"about_ca_topic_score_codex":0.00028701004,"about_ca_topic_score_gemma":0.0007427357,"teacher_disagreement_score":0.5365803,"about_ca_system_score_codex":0.00003900081,"about_ca_system_score_gemma":0.000095549636,"threshold_uncertainty_score":0.34753522},"labels":[],"label_agreement":null},{"id":"W4400382590","doi":"10.1007/s12561-024-09444-7","title":"A Comparison of Estimation Methods for Shared Gamma Frailty Models","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University; University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Estimator; Computer science; Estimation; Expectation–maximization algorithm; Maximization; Maximum likelihood; Series (stratigraphy); Data mining; Statistics; Mathematical optimization; Mathematics","score_opus":0.40930061629507997,"score_gpt":0.5920599909361398,"score_spread":0.18275937464105985,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400382590","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":[0.0009324996,0.00019869917,0.9962768,0.000065186614,0.00041115127,0.0002898593,0.0007995944,0.00003823184,0.0009879757],"genre_scores_gemma":[0.14675401,0.000008936987,0.85309523,0.000010068735,0.000013853274,0.000042384007,0.000009367351,0.0000087431135,0.00005742206],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.998555,0.00014264445,0.0005620616,0.00028936774,0.00023027694,0.00022064467],"domain_scores_gemma":[0.99140185,0.008172382,0.000111395275,0.00015047913,0.000112134956,0.000051753632],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0015759878,0.000117245676,0.00032970894,0.00015712247,0.000056883062,0.000085942425,0.00022854727,0.00006115339,0.0000858557],"category_scores_gemma":[0.0061226934,0.000093678405,0.00003435345,0.00043017522,0.00026591134,0.00012075339,0.00004166589,0.00010516938,0.0000017907316],"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":[0.0000063394505,0.00004567707,0.00005353154,0.00032472424,0.0000045210727,5.7294835e-7,0.00090887485,0.0001142265,0.00076451787,0.8117999,0.0008043616,0.18517277],"study_design_scores_gemma":[0.00004343872,0.00007970897,0.00010579772,0.00008368539,0.000011019973,3.450436e-7,0.00011339537,0.4909984,0.0014384057,0.5069896,0.00007520002,0.000061022864],"about_ca_topic_score_codex":0.000028316786,"about_ca_topic_score_gemma":0.000023694209,"teacher_disagreement_score":0.49088416,"about_ca_system_score_codex":0.000032111264,"about_ca_system_score_gemma":0.00011161897,"threshold_uncertainty_score":0.7329878},"labels":[],"label_agreement":null},{"id":"W4400617794","doi":"10.1007/s12561-024-09446-5","title":"Analyzing RNA-Seq Data in Complex Study Designs","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Single-cell and spatial transcriptomics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Fonds de recherche du Québec – Nature et technologies; National Institute of General Medical Sciences; U.S. Department of Defense","keywords":"Heteroscedasticity; Econometrics; Normalization (sociology); Biostatistics; Regression; Computer science; Statistics; Data mining; Mathematics; Medicine; Internal medicine","score_opus":0.12275928489149311,"score_gpt":0.36566508834239253,"score_spread":0.2429058034508994,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400617794","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":[0.92460334,0.00082695705,0.07228425,0.00009890459,0.0006166488,0.00027667213,0.0006040651,0.000018218443,0.00067094475],"genre_scores_gemma":[0.9883099,0.00014469451,0.01113958,0.000043326858,0.000063556836,0.000004741134,0.000204481,0.000008881959,0.000080824604],"study_design_codex":"bench_or_experimental","study_design_gemma":"observational","domain_scores_codex":[0.99877423,0.00007642111,0.00024621238,0.0005161147,0.00015660294,0.00023042187],"domain_scores_gemma":[0.9995318,0.00005141544,0.000023803497,0.0003319165,0.000019844445,0.000041224725],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00061477814,0.00010943597,0.00012029918,0.00013063276,0.000056628534,0.000119258904,0.00055649073,0.000045748497,0.000020863901],"category_scores_gemma":[0.00014083173,0.00009724541,0.000012936974,0.00040425174,0.00017065812,0.000009541679,0.00014320444,0.00010033365,0.000004745011],"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":[0.00005690199,0.0006911748,0.31305042,0.00009977413,0.000041743973,0.00021479916,0.0011937037,0.0002732266,0.65220565,0.002643747,0.003394374,0.026134502],"study_design_scores_gemma":[0.0040638493,0.00476085,0.5809706,0.00041434227,0.0001740789,0.000043761174,0.01071532,0.30849123,0.041054886,0.013624125,0.03246883,0.0032181214],"about_ca_topic_score_codex":0.0004988009,"about_ca_topic_score_gemma":0.006300528,"teacher_disagreement_score":0.61115074,"about_ca_system_score_codex":0.000016443022,"about_ca_system_score_gemma":0.00012142546,"threshold_uncertainty_score":0.39655527},"labels":[],"label_agreement":null},{"id":"W4403567836","doi":"10.1007/s12561-024-09461-6","title":"Benefits of Repeated Matched-Cohort and Nested Case–Control Analyses with Time-dependent Exposure in Observational Studies","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":2,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"BC Centre for Disease Control; St. Paul's Hospital; University of British Columbia","funders":"","keywords":"Observational study; Biostatistics; Cohort study; Econometrics; Medicine; Statistics; Epidemiology; Internal medicine; Mathematics","score_opus":0.24423021167723682,"score_gpt":0.45873918055149415,"score_spread":0.21450896887425733,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4403567836","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":[0.9725698,0.0014065572,0.024798598,0.00011980423,0.000054851982,0.0004063577,0.00044721682,0.00010154525,0.00009527263],"genre_scores_gemma":[0.8856024,0.00020037127,0.11404717,0.000015870113,0.000008278239,0.00003844966,0.0000054576853,0.000009792225,0.00007216392],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985832,0.00007988686,0.00047399814,0.00032856056,0.00034884815,0.00018548027],"domain_scores_gemma":[0.99771637,0.0017947151,0.00012909567,0.00015112844,0.00017705209,0.000031630396],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008222665,0.00015191404,0.00036144556,0.00027024004,0.0000439122,0.00004363909,0.00011296719,0.00005080055,0.000016407585],"category_scores_gemma":[0.0010695928,0.00010920611,0.000011897992,0.0006871636,0.00037431705,0.00021481531,0.00004580085,0.00012393681,7.739129e-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":[0.00010923398,0.00030738176,0.5056711,0.0012119192,0.00031885545,0.002243339,0.0063174493,0.0027538158,0.009129347,0.460805,0.00035135695,0.010781194],"study_design_scores_gemma":[0.0007983644,0.0009932114,0.21893974,0.0017890712,0.00027338037,0.00051678374,0.005142753,0.02117609,0.00921748,0.7404048,0.000008275336,0.0007400637],"about_ca_topic_score_codex":0.00010692405,"about_ca_topic_score_gemma":0.0015228419,"teacher_disagreement_score":0.28673136,"about_ca_system_score_codex":0.000060781116,"about_ca_system_score_gemma":0.000083741885,"threshold_uncertainty_score":0.44532958},"labels":[],"label_agreement":null},{"id":"W4404761272","doi":"10.1007/s12561-024-09467-0","title":"Generalized Linear Mixed Models with Censored Covariates and Measurement Errors, with Applications in HIV/AIDS Studies","year":2024,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"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","keywords":"Covariate; Biostatistics; Econometrics; Human immunodeficiency virus (HIV); Statistics; Generalized linear model; Medicine; Generalized linear mixed model; Computer science; Mathematics; Epidemiology; Virology; Internal medicine","score_opus":0.05359051196322061,"score_gpt":0.3201356550565645,"score_spread":0.2665451430933439,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4404761272","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":[0.0016188,0.0017038661,0.9955026,0.00045550466,0.000099113684,0.00039554262,0.000049705846,0.00006689442,0.000107999826],"genre_scores_gemma":[0.15538684,0.0002775884,0.8441247,0.00005319139,0.000011908385,0.00009579066,0.0000024961748,0.00000865574,0.00003885065],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9982922,0.00013444226,0.0002469883,0.00061870454,0.00043289643,0.000274732],"domain_scores_gemma":[0.99929917,0.00018370508,0.00005420797,0.00025710385,0.00013668195,0.00006913508],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011968648,0.00017107984,0.00025507296,0.0002183316,0.00010675684,0.00030246456,0.00057092455,0.000034183922,6.0253416e-7],"category_scores_gemma":[0.000044245917,0.00010615979,0.000008924579,0.0010892616,0.00034793562,0.0003949559,0.00018346519,0.00012163342,4.753629e-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":[0.000008159582,0.000034653116,0.00034323757,0.000044110824,0.000014270144,0.000016487284,0.0019574317,0.0005421229,0.00014365089,0.99113315,0.00009074636,0.005671966],"study_design_scores_gemma":[0.0004451474,0.00021318717,0.000592482,0.00020419955,0.000016024334,0.000009273945,0.0005144547,0.43415394,0.00041120406,0.5628208,0.00034518194,0.00027414758],"about_ca_topic_score_codex":0.00009489943,"about_ca_topic_score_gemma":0.00083181565,"teacher_disagreement_score":0.4336118,"about_ca_system_score_codex":0.000062536055,"about_ca_system_score_gemma":0.00021974486,"threshold_uncertainty_score":0.43290704},"labels":[],"label_agreement":null},{"id":"W4406133897","doi":"10.1007/s12561-024-09471-4","title":"Dynamic Treatment Regimes with Replicated Observations Available for Error-Prone Covariates: A Q-Learning Approach","year":2025,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":1,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Covariate; Biostatistics; Econometrics; Computer science; Statistics; Medicine; Mathematics; Epidemiology; Internal medicine","score_opus":0.14020212861599984,"score_gpt":0.39800893538283705,"score_spread":0.2578068067668372,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406133897","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":[0.004235925,0.000047639067,0.9875055,0.0002212063,0.00008786933,0.0007183254,0.0002958109,0.00006758154,0.0068201656],"genre_scores_gemma":[0.04140963,0.00003542662,0.95143515,0.000049115737,0.0000052236164,0.0003338105,0.000034769757,0.000011145053,0.0066857436],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99849147,0.000110430185,0.00037394103,0.0004982154,0.0001884755,0.00033749823],"domain_scores_gemma":[0.9962792,0.0030536863,0.0001573274,0.00028736496,0.00017040697,0.00005203212],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006606183,0.00017747117,0.0003154852,0.00014730233,0.0002827023,0.00009173708,0.00023268977,0.000061406856,0.000040528354],"category_scores_gemma":[0.004324141,0.00012534106,0.000021935386,0.0008049153,0.0003450445,0.000062505766,0.0000364285,0.00011093764,0.000003536492],"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":[0.000044067834,0.00021762309,0.002543609,0.00013068356,0.00002404221,0.0000018092546,0.00034406674,0.00009892949,0.00023629243,0.9922191,0.00094050745,0.0031992546],"study_design_scores_gemma":[0.00072522415,0.00070301606,0.004665027,0.00017519279,0.000078434816,0.0000021398805,0.0009723871,0.35492292,0.00018687356,0.63588345,0.0014208964,0.00026444308],"about_ca_topic_score_codex":0.0001612587,"about_ca_topic_score_gemma":0.00014452254,"teacher_disagreement_score":0.35633567,"about_ca_system_score_codex":0.00013011396,"about_ca_system_score_gemma":0.00028860662,"threshold_uncertainty_score":0.5176713},"labels":[],"label_agreement":null},{"id":"W4406834933","doi":"10.1007/s12561-026-09521-z","title":"Stability and Performance of Linear Combination Tests of Gene Set Enrichment for Multiple Covariance Estimators in Unbalanced Studies","year":2025,"lang":"en","type":"preprint","venue":"Statistics in Biosciences","topic":"Genetic Associations and Epidemiology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia; University of Alberta","funders":"","keywords":"Estimator; Covariance; Set (abstract data type); Stability (learning theory); Mathematics; Statistics; Econometrics; Computer science; Machine learning","score_opus":0.04449617047803959,"score_gpt":0.3631061786600419,"score_spread":0.3186100081820023,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406834933","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":[0.9837668,0.000843152,0.013094494,0.00005579084,0.00024752243,0.00041970462,0.001560294,0.0000015868751,0.000010671158],"genre_scores_gemma":[0.86370206,0.0013163319,0.13472858,0.000013364168,0.000007816723,0.000059858918,0.00015792876,0.0000023209634,0.000011738788],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9988181,0.000085546955,0.0005002709,0.000349836,0.00009370734,0.0001525186],"domain_scores_gemma":[0.9989091,0.00030627009,0.00034437442,0.00017886837,0.00024403923,0.00001735378],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00095065444,0.00011865573,0.00035333363,0.000082443556,0.00003739077,0.000003292592,0.00016096253,0.00013867083,5.069738e-7],"category_scores_gemma":[0.0024009082,0.00011204887,0.000021090571,0.00013039842,0.00034563348,0.0000020930574,0.00019368243,0.000075200274,2.6764873e-8],"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":[0.000051803967,0.00010339924,0.98371005,0.00067474425,0.000024655932,1.215722e-7,0.00037773294,0.00583474,0.008002072,0.00033281784,0.000060679005,0.0008272068],"study_design_scores_gemma":[0.00078582467,0.00056749175,0.875581,0.00016212245,0.000020043379,2.2124578e-7,0.00051238155,0.09557926,0.023000201,0.0035755825,0.000025880072,0.00018997496],"about_ca_topic_score_codex":0.00008651347,"about_ca_topic_score_gemma":0.00047433446,"teacher_disagreement_score":0.12163408,"about_ca_system_score_codex":0.00003505089,"about_ca_system_score_gemma":0.00021227756,"threshold_uncertainty_score":0.45692202},"labels":[],"label_agreement":null},{"id":"W4407888374","doi":"10.1007/s12561-025-09475-8","title":"Introduction to Special Issue on Machine Learning in Biomedical Sciences","year":2025,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Genetics, Bioinformatics, and Biomedical Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Biostatistics; Computer science; Data science; Artificial intelligence; Medicine; Public health; Pathology","score_opus":0.013366945041809413,"score_gpt":0.33398814331143883,"score_spread":0.3206211982696294,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4407888374","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":[0.796179,0.0006477359,0.046309978,0.06708642,0.013937007,0.0020971186,0.0005978646,0.00007341101,0.07307145],"genre_scores_gemma":[0.9288396,0.0013667891,0.049283303,0.0025917515,0.0079954825,0.000044223005,0.0002535166,0.000016166665,0.009609128],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99829614,0.00007886952,0.00032689524,0.00041760146,0.0004805764,0.00039989452],"domain_scores_gemma":[0.9995987,0.00006659302,0.000040745883,0.0001373914,0.000050497103,0.000106107436],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011792406,0.00011518748,0.00013446098,0.0005297228,0.00014307683,0.00007009975,0.0004085024,0.00009899736,0.00011523873],"category_scores_gemma":[0.0024144724,0.00009423434,0.000017939843,0.0010395533,0.0008578451,0.0000050914,0.000183807,0.00018970255,0.000047837933],"study_design_candidate":"not_applicable","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":[0.00041516422,0.00089212606,0.042777468,0.00020012737,0.000022046375,0.000029225896,0.0010446665,0.0012102267,0.08541878,0.015178962,0.2902145,0.5625967],"study_design_scores_gemma":[0.0006528845,0.0024274269,0.019313673,0.00007681718,0.000003453636,0.000002652531,0.0010261144,0.006111136,0.02015275,0.0023536703,0.9475644,0.00031502056],"about_ca_topic_score_codex":0.00008574386,"about_ca_topic_score_gemma":0.0007268173,"teacher_disagreement_score":0.6573499,"about_ca_system_score_codex":0.000038954462,"about_ca_system_score_gemma":0.00025426727,"threshold_uncertainty_score":0.38427648},"labels":[],"label_agreement":null},{"id":"W4411394384","doi":"10.1007/s12561-025-09495-4","title":"Causal Inference on Missing Exposure via Robust Estimation","year":2025,"lang":"en","type":"article","venue":"Statistics in Biosciences","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":0,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Causal inference; Inference; Biostatistics; Estimation; Missing data; Econometrics; Computer science; Statistics; Data mining; Artificial intelligence; Machine learning; Mathematics; Public health; Economics; Medicine","score_opus":0.12353098767810926,"score_gpt":0.44517602030833536,"score_spread":0.3216450326302261,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411394384","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":[0.009065073,0.000020298701,0.98527664,0.0002416872,0.00026111686,0.00021472452,0.00004209003,0.00015452274,0.004723838],"genre_scores_gemma":[0.52633625,0.000012141732,0.47336772,0.000101151956,0.000008177901,0.000016550372,0.0000047348894,0.0000051765396,0.00014808781],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985863,0.00007441611,0.00039148086,0.00033268196,0.0003244872,0.00029064328],"domain_scores_gemma":[0.997872,0.0015683766,0.0001414169,0.00027245417,0.00009916766,0.00004657794],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005121277,0.00016961063,0.00020998345,0.00029567714,0.00015760362,0.00009195153,0.00031695058,0.00008686268,0.000041265474],"category_scores_gemma":[0.0036018107,0.0001494926,0.000016001548,0.00065766694,0.00033335027,0.00019469652,0.0000700777,0.00021709816,0.000007843479],"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":[0.000009132985,0.00009545925,0.0019800398,0.00007116728,0.0000030252008,0.000011001465,0.0002463344,0.00061831035,0.0012388858,0.9434539,0.00084930554,0.05142341],"study_design_scores_gemma":[0.00010748349,0.00013566583,0.00255707,0.00025792734,0.00000703585,0.0000011095822,0.0001068896,0.0350729,0.005147425,0.9563825,0.000055789442,0.00016817133],"about_ca_topic_score_codex":0.000044550492,"about_ca_topic_score_gemma":0.00020383917,"teacher_disagreement_score":0.51727116,"about_ca_system_score_codex":0.00012575118,"about_ca_system_score_gemma":0.00014757129,"threshold_uncertainty_score":0.6096131},"labels":[],"label_agreement":null}]}