{"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0111632515,0.003616509,0.97731704,0.0041745505,0.0007121685,0.000119908254,0.0016109021,0.0003256586,0.0009600719],"genre_scores_gemma":[0.545195,0.009756204,0.40763092,0.0034647856,0.0038991943,0.0025341152,0.0041865394,0.00033481888,0.022998383],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9782004,0.015527117,0.00096460595,0.0025760827,0.001968759,0.0007631323],"domain_scores_gemma":[0.8556361,0.116129525,0.005900452,0.019128963,0.0024183958,0.00078647106],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04621983,0.0016620005,0.0045510107,0.0018128805,0.0010683861,0.0028208,0.0049483073,0.0039362297,0.0068266704],"category_scores_gemma":[0.15403526,0.0018701578,0.0023716076,0.004102486,0.0030736174,0.0058393436,0.0020626434,0.007043127,0.0013925014],"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":[0.0012925094,0.00017503754,0.016216356,0.0010604667,0.0015076743,0.0011354807,0.0005031897,0.044601273,0.00048384562,0.8257819,0.0155696785,0.091672555],"study_design_scores_gemma":[0.00035241776,0.00016414508,0.0028460799,0.00018052344,0.0006233665,0.00062450656,0.00006927298,0.13943289,0.0005092895,0.84491277,0.010205398,0.00007936477],"about_ca_topic_score_codex":0.0046847565,"about_ca_topic_score_gemma":0.0041742744,"teacher_disagreement_score":0.04621983,"about_ca_system_score_codex":0.0014239243,"about_ca_system_score_gemma":0.0051181074,"threshold_uncertainty_score":0.24443674},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0020924376,0.00014100879,0.9965056,0.00021045703,0.0000215251,0.000025232368,0.00009644149,0.00059521996,0.0003119711],"genre_scores_gemma":[0.21314412,0.0005462425,0.7810761,0.0005041546,0.00018944565,0.00038908183,0.0007857919,0.00055567175,0.0028094063],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9966972,0.0020242254,0.00017655818,0.00051127333,0.00047336332,0.0001174629],"domain_scores_gemma":[0.96668184,0.029085346,0.0005589903,0.0020837048,0.0011620603,0.00042796638],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009963545,0.0011523799,0.0018692849,0.0016431959,0.00073804933,0.0021107593,0.0035604783,0.001968005,0.007966569],"category_scores_gemma":[0.05269017,0.0010922925,0.0013940212,0.0022163475,0.0020091524,0.0052261287,0.0043671336,0.0047402456,0.0015224998],"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":[0.0005134556,0.00033963245,0.0025729325,0.00041402993,0.00017547423,0.00016455079,0.00024227638,0.3262294,0.0010765144,0.22294696,0.012984144,0.43234065],"study_design_scores_gemma":[0.000057040903,0.000028943443,0.00008589246,0.000020924823,0.000014301478,0.0000162198,0.000011609889,0.7791514,0.0004317581,0.21906306,0.0011093585,0.000009574675],"about_ca_topic_score_codex":0.002623636,"about_ca_topic_score_gemma":0.003363598,"teacher_disagreement_score":0.009963545,"about_ca_system_score_codex":0.0011630467,"about_ca_system_score_gemma":0.0023952776,"threshold_uncertainty_score":0.05269289},"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":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.046595667,0.017684247,0.9077765,0.0055889157,0.002947329,0.015659783,0.0008227761,0.0004822987,0.0024425038],"genre_scores_gemma":[0.4779322,0.007462329,0.4551355,0.003476669,0.0013080817,0.051584158,0.000711302,0.00011241077,0.0022774471],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.73327446,0.24902809,0.005089503,0.00864507,0.0030930012,0.0008697967],"domain_scores_gemma":[0.6571305,0.3144036,0.012891885,0.011238269,0.0022835718,0.0020521758],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.16551183,0.0033204372,0.011030828,0.0022228812,0.00095903117,0.0035086398,0.0038279698,0.007880926,0.0043372973],"category_scores_gemma":[0.30822337,0.0023470514,0.0048121726,0.0018385472,0.004984679,0.0033478753,0.0019635577,0.0066312896,0.00073218654],"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":[0.20034748,0.0044799205,0.009022453,0.030446526,0.020845402,0.0012033362,0.0013103856,0.09534984,0.003616559,0.30070737,0.00984644,0.3228242],"study_design_scores_gemma":[0.09176468,0.03744294,0.0031864403,0.0028919491,0.013551132,0.0008246311,0.00024876161,0.22116265,0.002556728,0.6141004,0.011942671,0.00032703683],"about_ca_topic_score_codex":0.00048370587,"about_ca_topic_score_gemma":0.00047804724,"teacher_disagreement_score":0.83448815,"about_ca_system_score_codex":0.0015909789,"about_ca_system_score_gemma":0.005542908,"threshold_uncertainty_score":0.8753207},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0044292514,0.000340207,0.993573,0.0003992623,0.00013543593,0.00013869265,0.0002760777,0.00026396863,0.00044403045],"genre_scores_gemma":[0.22013767,0.00093122176,0.7702382,0.0007879742,0.0007462595,0.002573471,0.0010784523,0.00013406812,0.0033727824],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9826166,0.014412904,0.00040793626,0.0012712891,0.0010572341,0.0002338875],"domain_scores_gemma":[0.8672448,0.116910815,0.002654762,0.010488719,0.002181555,0.00051931833],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.030427925,0.0015366788,0.0030975847,0.001753035,0.0009471149,0.0016389748,0.00405605,0.002506667,0.005409347],"category_scores_gemma":[0.09580054,0.0012590267,0.003558569,0.0022251594,0.0025793102,0.002608863,0.0029177205,0.0042521874,0.0006575167],"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":[0.0024940234,0.00066224,0.008974237,0.0008855794,0.0017017059,0.00042157993,0.00059105584,0.22982837,0.00269372,0.32906258,0.007819652,0.41486531],"study_design_scores_gemma":[0.0003345057,0.00037770515,0.0011441419,0.00013042812,0.00042543223,0.00018958113,0.00002978779,0.814077,0.001069308,0.17808448,0.0040712697,0.00006631957],"about_ca_topic_score_codex":0.0025592416,"about_ca_topic_score_gemma":0.0024084866,"teacher_disagreement_score":0.030427925,"about_ca_system_score_codex":0.00091342704,"about_ca_system_score_gemma":0.0029374072,"threshold_uncertainty_score":0.16092014},"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":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.14682092,0.00065012724,0.8434389,0.00024069076,0.000082267725,0.00010047175,0.0037297173,0.0033107593,0.0016261019],"genre_scores_gemma":[0.6983996,0.0003526745,0.28847197,0.00018010817,0.00016060269,0.0006167879,0.008448809,0.0007873984,0.0025821102],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99681133,0.0007894405,0.00029408032,0.00096726447,0.00090905564,0.00022880883],"domain_scores_gemma":[0.9847076,0.011770747,0.0009244028,0.0017210845,0.00068371405,0.0001924888],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005075667,0.00058103516,0.0014366742,0.0020612157,0.0006280425,0.001495884,0.0008497621,0.0004583651,0.002208289],"category_scores_gemma":[0.010980271,0.0002771951,0.0015497686,0.0030223657,0.0008147206,0.0013155983,0.00069493864,0.0013406223,0.0005588127],"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":[0.0025028964,0.00050937565,0.099436104,0.00074159144,0.0017223364,0.0005381759,0.00046790682,0.046882752,0.24447358,0.036573835,0.006163766,0.5599877],"study_design_scores_gemma":[0.0000986774,0.0011697727,0.12756877,0.000065449334,0.0008910312,0.0015704035,0.00043745636,0.6137091,0.17249598,0.0655114,0.016323896,0.00015805158],"about_ca_topic_score_codex":0.001831946,"about_ca_topic_score_gemma":0.0023218167,"teacher_disagreement_score":0.005075667,"about_ca_system_score_codex":0.000938192,"about_ca_system_score_gemma":0.0022117894,"threshold_uncertainty_score":0.026843011},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.056429364,0.00235379,0.93400127,0.0018799923,0.0009801498,0.0012386296,0.0010314953,0.00058026,0.0015050957],"genre_scores_gemma":[0.6850227,0.0013142327,0.29622173,0.0011517012,0.001278842,0.00776189,0.0014318958,0.00020231576,0.00561467],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8448589,0.12001525,0.0076718805,0.01730759,0.008341292,0.001805139],"domain_scores_gemma":[0.4480134,0.49175122,0.016415276,0.03867939,0.0036274607,0.0015132136],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1735333,0.002700772,0.0073166112,0.0018944481,0.0012211633,0.0025995243,0.0043043676,0.0038291665,0.010506467],"category_scores_gemma":[0.34008238,0.0018385472,0.00665056,0.0027099266,0.004520831,0.00555738,0.0029622626,0.0073567065,0.00083583425],"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":[0.023690075,0.0014528775,0.04410704,0.004261077,0.013452446,0.0030302305,0.0018478137,0.09715118,0.0022575657,0.29585335,0.008773391,0.50412303],"study_design_scores_gemma":[0.0035977704,0.0060788095,0.008421032,0.00044673565,0.0040903413,0.0017263327,0.00016639537,0.44904748,0.0026406,0.517911,0.0057325275,0.00014090797],"about_ca_topic_score_codex":0.0013424073,"about_ca_topic_score_gemma":0.0012029966,"teacher_disagreement_score":0.1735333,"about_ca_system_score_codex":0.0015741012,"about_ca_system_score_gemma":0.0051752673,"threshold_uncertainty_score":0.9177428},"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":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026018882,0.0011640835,0.9689172,0.0011378296,0.0003751525,0.00019442545,0.0013977141,0.00040440905,0.00039031162],"genre_scores_gemma":[0.46720362,0.002686867,0.51056635,0.0011383987,0.0010229718,0.004755603,0.0032783833,0.00037006979,0.008977696],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9816768,0.013529539,0.00090200885,0.0023651235,0.0010886578,0.0004377914],"domain_scores_gemma":[0.9041276,0.071719564,0.0049720705,0.016921466,0.0015425606,0.0007167809],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.04150794,0.0011532728,0.002245859,0.0013952332,0.0009933742,0.0014954944,0.0027511783,0.0019250473,0.006359917],"category_scores_gemma":[0.10586565,0.00085413887,0.0027494845,0.0026513152,0.0020299763,0.00283028,0.0023641225,0.003302529,0.0010315228],"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":[0.0046574175,0.0012958096,0.13318478,0.0018138579,0.0057600266,0.0012137244,0.0017057951,0.03285204,0.006976749,0.41335618,0.01496709,0.3822166],"study_design_scores_gemma":[0.00080782274,0.0030551238,0.044322174,0.00045357837,0.0031119199,0.0016173106,0.0005673711,0.25048545,0.005757014,0.6636168,0.02604919,0.00015626055],"about_ca_topic_score_codex":0.0019708017,"about_ca_topic_score_gemma":0.0012867391,"teacher_disagreement_score":0.95849204,"about_ca_system_score_codex":0.00074710115,"about_ca_system_score_gemma":0.0038883053,"threshold_uncertainty_score":0.21951759},"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":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010593138,0.013594959,0.93276674,0.03995336,0.0010909188,0.000105702646,0.00041281438,0.0000934116,0.0013889389],"genre_scores_gemma":[0.4600145,0.020044371,0.4885615,0.013070836,0.011117306,0.0013422755,0.0007205532,0.0002945714,0.0048340443],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.88939536,0.09873957,0.0035256646,0.0040441225,0.0035741213,0.00072124775],"domain_scores_gemma":[0.40294385,0.5743007,0.0064698663,0.010726721,0.0043821936,0.0011766595],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.14910667,0.0011278732,0.0028645983,0.0025034547,0.0020920536,0.0053055924,0.005314206,0.0059626373,0.0024285365],"category_scores_gemma":[0.40879583,0.0016248464,0.0030719186,0.0047679874,0.008293853,0.0069480417,0.0033560982,0.012150306,0.00029106686],"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":[0.00010454693,0.00010688298,0.012521087,0.00046872025,0.00085789856,0.00052944303,0.0010193755,0.036609653,0.00011856461,0.90450585,0.005834013,0.03732401],"study_design_scores_gemma":[0.00004383108,0.000036762103,0.0014496642,0.00015003853,0.00015964969,0.0001567215,0.00013128018,0.058591083,0.00008233814,0.9351676,0.004003249,0.000027676628],"about_ca_topic_score_codex":0.020462872,"about_ca_topic_score_gemma":0.011396146,"teacher_disagreement_score":0.8508933,"about_ca_system_score_codex":0.0031842585,"about_ca_system_score_gemma":0.006620217,"threshold_uncertainty_score":0.78856087},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00049342745,0.00057944393,0.9983754,0.00017454718,0.0000648578,0.00003493808,0.00003392126,0.000086615604,0.00015679093],"genre_scores_gemma":[0.017115192,0.0023567786,0.9770981,0.0002819803,0.0006145709,0.0006998184,0.00021271476,0.00034365026,0.0012772395],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.97986335,0.017213691,0.0006125395,0.00096272206,0.0010865057,0.00026115033],"domain_scores_gemma":[0.8550345,0.13754636,0.0017798307,0.0030814838,0.0021218208,0.00043597075],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.033099223,0.0030769065,0.0039195623,0.0038326236,0.0012310364,0.0019129389,0.0055408217,0.0039588767,0.004351051],"category_scores_gemma":[0.10896009,0.0023325912,0.004082275,0.0043616407,0.0039159916,0.0033316065,0.0051789144,0.007562826,0.0013156271],"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":[0.00028619636,0.00027615254,0.002838346,0.00096747116,0.0008562676,0.0008562541,0.001234114,0.1656452,0.0029014961,0.63297683,0.009307439,0.18185422],"study_design_scores_gemma":[0.00009185237,0.00012332771,0.00092638,0.00017604526,0.00017954141,0.00032511368,0.000094240124,0.56515217,0.001161094,0.42236143,0.009211023,0.00019783073],"about_ca_topic_score_codex":0.007261128,"about_ca_topic_score_gemma":0.0074910424,"teacher_disagreement_score":0.033099223,"about_ca_system_score_codex":0.0018828292,"about_ca_system_score_gemma":0.0035785662,"threshold_uncertainty_score":0.17504752},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008171233,0.0021535654,0.98299384,0.0050054356,0.000212105,0.00019887264,0.00017611479,0.000108789005,0.0009799998],"genre_scores_gemma":[0.4840271,0.0060861968,0.49318302,0.0054504154,0.0034758823,0.0026808714,0.00084536104,0.00044710684,0.0038040942],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.7043076,0.25771955,0.009246737,0.01922306,0.007929209,0.0015738998],"domain_scores_gemma":[0.23979749,0.71493876,0.010300449,0.02975413,0.0043348214,0.0008743703],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.3258651,0.0025158226,0.010446531,0.004883107,0.0035922844,0.008857546,0.015537052,0.010578816,0.003836449],"category_scores_gemma":[0.6371128,0.004407018,0.0064006057,0.0069321515,0.027654584,0.016674574,0.006986284,0.018110123,0.00054161367],"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":[0.00041109588,0.00015575624,0.007961159,0.00073778676,0.0020725138,0.0011476264,0.0016929099,0.060392443,0.0004002864,0.8784742,0.00307692,0.043477334],"study_design_scores_gemma":[0.00012801745,0.00004259854,0.0017562485,0.00016878943,0.0002626151,0.00043931842,0.000119032826,0.0967073,0.00017193836,0.89882934,0.0012865118,0.00008822844],"about_ca_topic_score_codex":0.012218487,"about_ca_topic_score_gemma":0.0072773313,"teacher_disagreement_score":0.3258651,"about_ca_system_score_codex":0.003910393,"about_ca_system_score_gemma":0.004022354,"threshold_uncertainty_score":0.8313284},"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":"methods","genre_gemma":"methods","domain_codex":"methods","domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":"methods","prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.113193616,0.026842648,0.82895285,0.012625496,0.0020349377,0.001262044,0.0017766469,0.00078712846,0.012524562],"genre_scores_gemma":[0.89531404,0.002276343,0.09452489,0.0024824678,0.0014058894,0.0018125422,0.0004786983,0.00043484522,0.0012704054],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.4643305,0.4703803,0.019185757,0.026547689,0.017367816,0.0021879468],"domain_scores_gemma":[0.05351429,0.9163605,0.00432358,0.022721253,0.002293956,0.00078637974],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.37570614,0.0022967283,0.007926785,0.0068773455,0.002411589,0.0048369947,0.0056171967,0.008360936,0.006496626],"category_scores_gemma":[0.7515395,0.0029209491,0.0076402356,0.005322349,0.015317994,0.013832165,0.006889693,0.006976083,0.00076471386],"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":[0.008795169,0.0005225965,0.30172646,0.0075619216,0.034802742,0.0054303957,0.007443606,0.033063337,0.0043729045,0.35930815,0.012045528,0.22492714],"study_design_scores_gemma":[0.0026831075,0.0020347054,0.10178982,0.0012215283,0.008705775,0.0062370645,0.0010147427,0.07490934,0.003846163,0.7846326,0.012613842,0.00031132324],"about_ca_topic_score_codex":0.0010417412,"about_ca_topic_score_gemma":0.0008118334,"teacher_disagreement_score":0.62429386,"about_ca_system_score_codex":0.0012046902,"about_ca_system_score_gemma":0.0024748764,"threshold_uncertainty_score":0.7698655},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.46483782,0.0014469635,0.5188947,0.0009373458,0.0002647646,0.0005300488,0.003424938,0.0028084407,0.0068550585],"genre_scores_gemma":[0.71007663,0.00057214126,0.28332072,0.00017130436,0.00012475014,0.00019742546,0.0035843675,0.000080852114,0.0018717893],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99977607,0.000048742193,0.00001906865,0.00003673574,0.000080628364,0.00003868141],"domain_scores_gemma":[0.999468,0.00015342011,0.00008394351,0.00008732141,0.00016321382,0.00004399116],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00052281003,0.0004274303,0.0003110221,0.0022421062,0.00019792115,0.00045178327,0.00041107263,0.00044983765,0.0012382151],"category_scores_gemma":[0.0011723221,0.0001488646,0.00035478894,0.00141369,0.00021024041,0.00039388053,0.00030112438,0.00035598918,0.0004957311],"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":[0.0003536116,0.0006375227,0.020636708,0.00025228065,0.00008596122,0.00033967957,0.00013545323,0.033754777,0.14386924,0.002121701,0.010427625,0.7873854],"study_design_scores_gemma":[0.000052964944,0.00022467111,0.12031863,0.00007028395,0.00007681186,0.00029837893,0.00038398756,0.81705964,0.049033027,0.0041152486,0.00830517,0.00006124992],"about_ca_topic_score_codex":0.003419451,"about_ca_topic_score_gemma":0.006303833,"teacher_disagreement_score":0.003419451,"about_ca_system_score_codex":0.00031227028,"about_ca_system_score_gemma":0.0004394293,"threshold_uncertainty_score":0.006799102},"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":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.013295877,0.006780431,0.9733138,0.003099892,0.00029969277,0.0004361076,0.0008509908,0.0002795152,0.001643769],"genre_scores_gemma":[0.6582479,0.010352669,0.31487063,0.0033509606,0.0012811313,0.0032128228,0.0016343384,0.00052181626,0.006527619],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8464191,0.13917494,0.0038144798,0.0056452365,0.003639672,0.0013066264],"domain_scores_gemma":[0.2733433,0.70406467,0.007043681,0.01178123,0.0029625762,0.0008045391],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.18066321,0.0024354928,0.0079677785,0.0050098132,0.0015825861,0.005259702,0.006367319,0.0058812667,0.008346411],"category_scores_gemma":[0.48258272,0.0032465886,0.007099409,0.0036975911,0.004675234,0.006453524,0.0072338185,0.009042949,0.0004928666],"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":[0.0016517832,0.00040397723,0.0135331275,0.0043512033,0.01282073,0.0022556358,0.0011415733,0.4413422,0.0009073885,0.44390595,0.0071239034,0.070562504],"study_design_scores_gemma":[0.00021454193,0.00014468782,0.0023410576,0.0005099169,0.0016976041,0.00043754422,0.000120627126,0.37049004,0.0003875416,0.62031686,0.0032228983,0.000116669224],"about_ca_topic_score_codex":0.010719086,"about_ca_topic_score_gemma":0.0051841843,"teacher_disagreement_score":0.8193368,"about_ca_system_score_codex":0.0040498506,"about_ca_system_score_gemma":0.003880482,"threshold_uncertainty_score":0.9554498},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.029514685,0.000750206,0.96618515,0.0004660468,0.00007402559,0.0005065707,0.001535975,0.00029304926,0.00067429646],"genre_scores_gemma":[0.26935992,0.0028598004,0.7079367,0.00044791933,0.00018224167,0.0029947949,0.007252828,0.00014815111,0.008817794],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.98206306,0.013905076,0.00090493396,0.0017375355,0.0009769592,0.0004124319],"domain_scores_gemma":[0.8500655,0.1299514,0.008151916,0.0075101154,0.003820748,0.000500198],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.044247475,0.0014708604,0.0029739817,0.0024207376,0.0007380641,0.0022305949,0.003872252,0.0021339394,0.006078104],"category_scores_gemma":[0.107493095,0.0013297779,0.003656343,0.0043925364,0.0012632181,0.0019910338,0.0017976415,0.0036600272,0.0012575255],"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":[0.00035505992,0.00026897402,0.11659299,0.0010965707,0.0029782436,0.0014060554,0.002304672,0.41633007,0.0011296917,0.26781425,0.008342401,0.18138105],"study_design_scores_gemma":[0.00011635979,0.00023964536,0.017515825,0.00024193367,0.00056795747,0.00044475292,0.00034126008,0.8818047,0.00053234067,0.09179596,0.0062907115,0.00010852133],"about_ca_topic_score_codex":0.034672104,"about_ca_topic_score_gemma":0.03393705,"teacher_disagreement_score":0.044247475,"about_ca_system_score_codex":0.0021277536,"about_ca_system_score_gemma":0.0027328935,"threshold_uncertainty_score":0.23400581},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.006735734,0.0010518203,0.99047756,0.0008526987,0.00008179683,0.0000671029,0.00013004775,0.000097768774,0.0005055584],"genre_scores_gemma":[0.4329345,0.004153867,0.55663526,0.0008539005,0.0009830537,0.0010210781,0.0007692389,0.00012491143,0.0025241694],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9410753,0.047118455,0.002311138,0.004867392,0.0038303556,0.0007973863],"domain_scores_gemma":[0.5552384,0.41496345,0.009852179,0.01619756,0.0030911877,0.0006571079],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.08812229,0.001829097,0.0046694996,0.004069067,0.00091466424,0.0042366576,0.004957468,0.003636645,0.002856519],"category_scores_gemma":[0.3403566,0.0017237294,0.0039033492,0.00453838,0.0049027083,0.0063473308,0.0048858393,0.006474725,0.00033761925],"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":[0.00023492846,0.00018906435,0.008835635,0.00093788735,0.001984436,0.00041438182,0.00055207167,0.059577744,0.00047191928,0.853229,0.001285686,0.07228719],"study_design_scores_gemma":[0.00010303569,0.00009705318,0.0015834939,0.00015756726,0.00049844995,0.00014416102,0.00006163518,0.15032102,0.00051962025,0.84475493,0.001727812,0.000031224437],"about_ca_topic_score_codex":0.003144837,"about_ca_topic_score_gemma":0.0015876937,"teacher_disagreement_score":0.9118777,"about_ca_system_score_codex":0.0020554643,"about_ca_system_score_gemma":0.005058623,"threshold_uncertainty_score":0.4660408},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00013568978,0.99635774,0.0022121228,0.00035443803,0.00031789197,0.000026171296,0.000300351,0.00003466456,0.0002609786],"genre_scores_gemma":[0.00116105,0.9930402,0.0038045577,0.0006443789,0.0006305303,0.000070161885,0.00045660828,0.00004172762,0.00015075726],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99086285,0.0029497647,0.0027168363,0.0011839432,0.0021278092,0.0001588226],"domain_scores_gemma":[0.92351794,0.066408776,0.003641494,0.0011585403,0.005027478,0.00024581104],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014464421,0.002145636,0.006412018,0.009671901,0.00041817556,0.0024173555,0.0028739336,0.0018077287,0.0037983367],"category_scores_gemma":[0.055255685,0.001060411,0.004538158,0.0129515305,0.0016414017,0.0022159668,0.0013145205,0.0027448286,0.0016893817],"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":[0.00017933626,0.000060344584,0.001173346,0.11183415,0.0019594776,0.00006785559,0.000088450244,0.00070184836,0.0004922141,0.0026076308,0.025450403,0.8553849],"study_design_scores_gemma":[0.00021938056,0.0006107365,0.017501775,0.1889487,0.01422939,0.0019132778,0.00026289877,0.002081922,0.002373005,0.017358297,0.7540993,0.00040136566],"about_ca_topic_score_codex":0.0053624134,"about_ca_topic_score_gemma":0.0072836345,"teacher_disagreement_score":0.014464421,"about_ca_system_score_codex":0.0018419157,"about_ca_system_score_gemma":0.0075818216,"threshold_uncertainty_score":0.076496124},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.022942796,0.00085035013,0.97331494,0.0007677919,0.00009104429,0.00006949549,0.00046309605,0.00036296673,0.0011375812],"genre_scores_gemma":[0.7277022,0.0036464287,0.23667492,0.00076510455,0.000750153,0.0011491441,0.0026331102,0.00057948154,0.026099484],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99487585,0.0024990572,0.00028221993,0.0011292114,0.00072204904,0.000491553],"domain_scores_gemma":[0.9616149,0.030281324,0.0029482557,0.0023836265,0.001921913,0.0008499641],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013950063,0.0022586156,0.005075927,0.004472289,0.0010187645,0.0040478986,0.007529517,0.0051000626,0.006743441],"category_scores_gemma":[0.03917533,0.002420177,0.0041152514,0.003725949,0.0038482903,0.007000086,0.0030605101,0.0052793855,0.0017683703],"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":[0.0005015739,0.00017413835,0.0058814,0.00039357648,0.0004999797,0.00054728263,0.00065525365,0.42701104,0.0022861648,0.5345489,0.0032649743,0.02423576],"study_design_scores_gemma":[0.000070713504,0.00006159353,0.0011811301,0.000068996094,0.00017551318,0.00022412692,0.000046563793,0.85129136,0.00025155855,0.14558429,0.0009539285,0.00009019588],"about_ca_topic_score_codex":0.0074591283,"about_ca_topic_score_gemma":0.0051343013,"teacher_disagreement_score":0.013950063,"about_ca_system_score_codex":0.0019291696,"about_ca_system_score_gemma":0.0016789832,"threshold_uncertainty_score":0.07377589},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01542477,0.0006132584,0.98123884,0.0009473698,0.000062365616,0.000043405424,0.0001381478,0.00011083107,0.001421008],"genre_scores_gemma":[0.5949364,0.002131825,0.3992334,0.00041053965,0.00023680316,0.00034283954,0.0003612854,0.00006048921,0.0022864426],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99381363,0.00467126,0.000238747,0.0005441049,0.00059588044,0.00013630756],"domain_scores_gemma":[0.98035276,0.016448328,0.0013779913,0.0012751692,0.00034449092,0.00020121493],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007750229,0.00077828136,0.0009971875,0.001607863,0.0005540443,0.0021166147,0.0019600817,0.002017773,0.0011743099],"category_scores_gemma":[0.030102858,0.0009092197,0.0018085644,0.001986891,0.0021632842,0.0033481473,0.0035669908,0.0021646372,0.0002250429],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00004813135,0.00012726206,0.010800731,0.00018160338,0.00028593445,0.0001528494,0.00035318919,0.814444,0.0007023369,0.14750631,0.00038277038,0.025014902],"study_design_scores_gemma":[0.000015886506,0.000054614113,0.0022169675,0.000039559523,0.0000453823,0.000073781426,0.00008156554,0.7869318,0.00019414342,0.20841558,0.001894358,0.00003647224],"about_ca_topic_score_codex":0.014443345,"about_ca_topic_score_gemma":0.018751688,"teacher_disagreement_score":0.014443345,"about_ca_system_score_codex":0.0017729232,"about_ca_system_score_gemma":0.002051995,"threshold_uncertainty_score":0.04098761},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0051531736,0.00015140284,0.9918331,0.00023677212,0.00015479546,0.00005577321,0.00023229397,0.00024556482,0.0019369965],"genre_scores_gemma":[0.19365354,0.00078255543,0.78966975,0.00041209583,0.00060001144,0.00052095484,0.0021485938,0.00038969613,0.011822868],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9983588,0.0007818019,0.00006999066,0.00026478566,0.00037347354,0.00015115837],"domain_scores_gemma":[0.99688727,0.0011767974,0.00024901688,0.001019991,0.0004844483,0.00018241828],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016838562,0.0012164777,0.0011006645,0.00067724905,0.0005999111,0.0011837453,0.001189727,0.0010241147,0.0055087595],"category_scores_gemma":[0.00733575,0.000526152,0.0014793602,0.0010667219,0.0011646831,0.0019496973,0.0018910842,0.002162154,0.0023796253],"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":[0.0003796961,0.0003258587,0.001057809,0.00038696217,0.00019030397,0.00040681445,0.00024899704,0.1373857,0.016055962,0.4703955,0.04192585,0.33124053],"study_design_scores_gemma":[0.00004296291,0.000095487245,0.0004128668,0.000032342607,0.000028833783,0.00017084798,0.000057407615,0.6771978,0.0025469884,0.30998355,0.00939349,0.000037370526],"about_ca_topic_score_codex":0.002192634,"about_ca_topic_score_gemma":0.0033200067,"teacher_disagreement_score":0.0055087595,"about_ca_system_score_codex":0.0002763051,"about_ca_system_score_gemma":0.0013504976,"threshold_uncertainty_score":0.018428624},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.87600625,0.0072035617,0.03881598,0.044587582,0.00090052385,0.0003272868,0.0010005803,0.00006859797,0.031089695],"genre_scores_gemma":[0.9900435,0.000972507,0.0061452384,0.001011094,0.00015869172,0.000062605104,0.00009336624,0.00002708895,0.001485764],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.98320466,0.009906074,0.0018248812,0.0013544905,0.0024495411,0.0012603279],"domain_scores_gemma":[0.8774646,0.08067724,0.022256674,0.004395147,0.013084698,0.002121719],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.03514004,0.00036843997,0.0006424666,0.0016033885,0.0024711392,0.002487902,0.002496457,0.0024696551,0.004427228],"category_scores_gemma":[0.109744705,0.00039797567,0.0012034309,0.0027243712,0.002570072,0.0028922874,0.0010420281,0.0023908392,0.00035398992],"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":[0.0034421694,0.0008053947,0.7817699,0.001513046,0.00031265183,0.008084433,0.019335425,0.003957521,0.0058370763,0.06955715,0.0083104,0.09707488],"study_design_scores_gemma":[0.00015400376,0.0054783532,0.7728739,0.0017197632,0.00079695275,0.008498061,0.07499713,0.00874578,0.008007453,0.059949815,0.058490798,0.00028789174],"about_ca_topic_score_codex":0.015299765,"about_ca_topic_score_gemma":0.023812538,"teacher_disagreement_score":0.03514004,"about_ca_system_score_codex":0.0034914813,"about_ca_system_score_gemma":0.0043605994,"threshold_uncertainty_score":0.18584049},"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":"editorial","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00015134827,0.0005450133,0.0012430386,0.10621308,0.88511497,0.000035253124,0.0030113726,0.0003522974,0.0033336368],"genre_scores_gemma":[0.016288504,0.0053233504,0.014520339,0.26570246,0.38109493,0.00065342494,0.0053424984,0.003976149,0.30709833],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9911327,0.0020795565,0.0019175649,0.0010658107,0.0032292767,0.0005750865],"domain_scores_gemma":[0.9009815,0.035822723,0.0050376668,0.0046627875,0.05049107,0.0030042839],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009719233,0.0023415333,0.0022863636,0.0044604368,0.0038890995,0.0050654775,0.004521788,0.0075763627,0.051392585],"category_scores_gemma":[0.16436797,0.001435638,0.0017318206,0.004702285,0.0033583161,0.0029923138,0.0020440628,0.015261565,0.037239976],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000016246662,0.0000037336615,0.000050274673,0.000053525528,0.0000036836898,0.000060765444,0.000024068395,0.00003114171,0.0000065117392,0.0005599769,0.9973411,0.0018491086],"study_design_scores_gemma":[0.000053180756,0.000019763598,0.0008755745,0.00049334246,0.000029538833,0.00025141911,0.00021548457,0.00039359296,0.00012961686,0.0025734515,0.99492043,0.000044670072],"about_ca_topic_score_codex":0.07527776,"about_ca_topic_score_gemma":0.067128494,"teacher_disagreement_score":0.07527776,"about_ca_system_score_codex":0.006931499,"about_ca_system_score_gemma":0.01175465,"threshold_uncertainty_score":0.1719253},"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":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1850715,0.0012282557,0.7896339,0.005064227,0.0007116531,0.0018362101,0.0013044007,0.004203648,0.010946155],"genre_scores_gemma":[0.873204,0.00021894865,0.11721531,0.0013534722,0.00019460458,0.0016883541,0.0005816301,0.00016017878,0.005383392],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99627715,0.0023967563,0.00009600373,0.0006477357,0.00034365163,0.00023865822],"domain_scores_gemma":[0.9938719,0.0048111863,0.00038824024,0.00042391868,0.00021349824,0.00029120673],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.011403675,0.00095990463,0.0015006935,0.00062671927,0.0003971071,0.0016549707,0.001024424,0.0017992262,0.007207683],"category_scores_gemma":[0.01266217,0.0004659176,0.0012339137,0.0003530362,0.0009268757,0.0009863032,0.001176511,0.0022700583,0.0008453988],"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":[0.08061194,0.004063404,0.010011859,0.0008895352,0.0023284063,0.0004522724,0.000233086,0.26016524,0.013639378,0.05162961,0.015801644,0.5601737],"study_design_scores_gemma":[0.012894037,0.0069703045,0.0022019595,0.000076515935,0.0012457717,0.000116146744,0.00004213189,0.91770816,0.007893506,0.042060327,0.00872003,0.000071033974],"about_ca_topic_score_codex":0.0010823822,"about_ca_topic_score_gemma":0.0010849419,"teacher_disagreement_score":0.9885963,"about_ca_system_score_codex":0.00059961644,"about_ca_system_score_gemma":0.002322453,"threshold_uncertainty_score":0.06030917},"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":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.009029052,0.00042060655,0.98908,0.00017581209,0.000052831576,0.0006552652,0.00010016936,0.00021867234,0.00026757419],"genre_scores_gemma":[0.0883312,0.00034135501,0.90718937,0.0001900195,0.00006836196,0.0030536063,0.0002788786,0.00014095998,0.00040617556],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.59578997,0.3844157,0.0046727527,0.006523788,0.007600462,0.0009974468],"domain_scores_gemma":[0.23253909,0.7360655,0.008146002,0.01435013,0.0075618066,0.0013374689],"candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.27439848,0.002736665,0.004828193,0.0027841337,0.0012239662,0.0028891894,0.0045927996,0.0051221857,0.0048488965],"category_scores_gemma":[0.51320946,0.0024480773,0.004351043,0.001964208,0.0037106923,0.004694146,0.0038403084,0.004027999,0.00063331553],"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":[0.018216934,0.0016611416,0.009601997,0.0037742602,0.0041086683,0.00026651187,0.001352243,0.35037258,0.004382939,0.20614673,0.0026145217,0.39750153],"study_design_scores_gemma":[0.004334723,0.0048881327,0.002949732,0.0005441017,0.0011382955,0.00019155002,0.000119959965,0.8640748,0.0038943458,0.11474443,0.002949708,0.00017017877],"about_ca_topic_score_codex":0.0017486421,"about_ca_topic_score_gemma":0.0011193953,"teacher_disagreement_score":0.72560155,"about_ca_system_score_codex":0.002934469,"about_ca_system_score_gemma":0.00729651,"threshold_uncertainty_score":0.8947959},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004960907,0.0010829278,0.9927584,0.0005354147,0.00007280533,0.000012359383,0.000044992954,0.00014283072,0.0003893539],"genre_scores_gemma":[0.3769907,0.004483614,0.60641444,0.00065851683,0.0010910868,0.00041980267,0.00066471635,0.0005883303,0.008688829],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9905043,0.0078004166,0.00023694296,0.00062023936,0.0006321977,0.00020584214],"domain_scores_gemma":[0.9525097,0.04299189,0.0011990034,0.0014837782,0.0015141264,0.0003015325],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.014679557,0.0011065945,0.0019189279,0.0016785319,0.0007807126,0.0017322452,0.002228907,0.0014039631,0.0023774088],"category_scores_gemma":[0.051562924,0.0009180807,0.0015962367,0.0022940396,0.0028359971,0.002378891,0.002060899,0.002684489,0.0006366622],"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":[0.00026683472,0.00011091791,0.0056980085,0.0007543374,0.00041568818,0.00031678614,0.000350115,0.23446377,0.0018500957,0.51326805,0.007122141,0.23538327],"study_design_scores_gemma":[0.0000335475,0.00004760405,0.0009211097,0.000057000685,0.0000535263,0.00007719727,0.000029285693,0.75944877,0.0005971965,0.23649031,0.002219343,0.000025179264],"about_ca_topic_score_codex":0.004808881,"about_ca_topic_score_gemma":0.0033434138,"teacher_disagreement_score":0.014679557,"about_ca_system_score_codex":0.00087242905,"about_ca_system_score_gemma":0.0019322416,"threshold_uncertainty_score":0.0776338},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0038260785,0.00015494983,0.9099637,0.004705874,0.0007550975,0.0006405688,0.003437917,0.06905465,0.0074610705],"genre_scores_gemma":[0.020354342,0.00023338197,0.95739514,0.001768649,0.00029219245,0.001252329,0.0027834747,0.009769355,0.0061510187],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99588364,0.0020265179,0.00033224683,0.00073124113,0.0008733626,0.00015283318],"domain_scores_gemma":[0.9546226,0.03445516,0.0009852096,0.003923332,0.004089761,0.0019240137],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0116366455,0.0022632084,0.0014858291,0.0022177694,0.0010966152,0.0047240765,0.0048372173,0.0019697805,0.08585914],"category_scores_gemma":[0.06945091,0.0016441136,0.0025404633,0.0016251588,0.0014889452,0.00478762,0.0031349815,0.0053412896,0.051070716],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007161907,0.0005189195,0.0047259443,0.0011270434,0.0002765388,0.0011951452,0.0016735485,0.016154546,0.008923995,0.016463507,0.49973658,0.448488],"study_design_scores_gemma":[0.0014652922,0.00043964342,0.011557594,0.001114946,0.00022283902,0.0015475478,0.0019557166,0.2548532,0.026383096,0.26551613,0.43421018,0.00073380437],"about_ca_topic_score_codex":0.0028256264,"about_ca_topic_score_gemma":0.0033348287,"teacher_disagreement_score":0.08585914,"about_ca_system_score_codex":0.00096871116,"about_ca_system_score_gemma":0.0017752951,"threshold_uncertainty_score":0.28722745},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.026198259,0.0005792416,0.96543235,0.004530505,0.00007575177,0.00007553024,0.00020177173,0.00017464656,0.0027319784],"genre_scores_gemma":[0.77770674,0.0012255403,0.21440117,0.0008763747,0.00041368903,0.00033700012,0.0005266863,0.00015066419,0.004362206],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.996512,0.0023450356,0.000121828925,0.0004560878,0.00029338256,0.00027158926],"domain_scores_gemma":[0.9296956,0.06413454,0.0022770807,0.0011440254,0.0019027137,0.00084607524],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.011942482,0.0008420793,0.0017845562,0.0013167929,0.00068924756,0.0026574754,0.0019090471,0.0020912506,0.0048883785],"category_scores_gemma":[0.0558853,0.00069094537,0.00084968156,0.0008570767,0.002204364,0.0038022955,0.0020033398,0.0033615367,0.00036921038],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009405141,0.000113991206,0.0034052331,0.00017174348,0.00007193903,0.00006054617,0.00016177657,0.69896287,0.00026118418,0.26093635,0.0035349608,0.032225337],"study_design_scores_gemma":[0.000013026542,0.000020093526,0.00018787573,0.000017670285,0.000006066727,0.0000060384377,0.000022397357,0.84987336,0.000080271006,0.14932229,0.0004446041,0.0000062155787],"about_ca_topic_score_codex":0.01698773,"about_ca_topic_score_gemma":0.009144374,"teacher_disagreement_score":0.01698773,"about_ca_system_score_codex":0.0030218617,"about_ca_system_score_gemma":0.0071003875,"threshold_uncertainty_score":0.06315863},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.98575526,0.0015505058,0.0017557908,0.00042253666,0.00005466289,0.000108066066,0.009786643,0.00005536508,0.0005112133],"genre_scores_gemma":[0.99168444,0.00043350342,0.0015344683,0.00008868603,0.000049785278,0.00006741764,0.005985369,0.000014249555,0.00014211306],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9717388,0.01753289,0.0033268484,0.003904366,0.0024114219,0.0010856425],"domain_scores_gemma":[0.86945003,0.109591864,0.010889019,0.0058162506,0.0026271406,0.0016257123],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.025946174,0.0009975253,0.001350639,0.0034074944,0.00054063014,0.002316824,0.0016161697,0.0010660432,0.0014939999],"category_scores_gemma":[0.08367671,0.00058635336,0.007684538,0.00580973,0.00054080505,0.0021924442,0.0020369086,0.0014568571,0.00018087971],"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":[0.0021323755,0.00024960196,0.97689724,0.0001761058,0.010119374,0.00011232215,0.00007206746,0.0030674988,0.00011765885,0.00015536656,0.00034551055,0.006554769],"study_design_scores_gemma":[0.00023807927,0.0014307621,0.9637024,0.00010032066,0.011402626,0.00019911781,0.00041631082,0.020469451,0.00052047294,0.0005670867,0.0009089917,0.000044347868],"about_ca_topic_score_codex":0.029825596,"about_ca_topic_score_gemma":0.0237045,"teacher_disagreement_score":0.029825596,"about_ca_system_score_codex":0.0010998914,"about_ca_system_score_gemma":0.0024903764,"threshold_uncertainty_score":0.13721812},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007408968,0.00019431714,0.9915895,0.00013072438,0.000023140066,0.000073166826,0.00006856987,0.0002215141,0.00029010186],"genre_scores_gemma":[0.2315317,0.0007257931,0.7656191,0.00017923432,0.0000872124,0.00043852336,0.00033679974,0.00018014273,0.00090147497],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9738584,0.021955183,0.0007643176,0.0013071307,0.0018271323,0.00028770466],"domain_scores_gemma":[0.9171844,0.0662945,0.0047925278,0.008871874,0.0022421696,0.0006144523],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.026864346,0.0013669059,0.0018548763,0.0048111975,0.00062697165,0.002424003,0.0018622476,0.0014362944,0.0033925842],"category_scores_gemma":[0.13343076,0.00068053976,0.002733044,0.0047602914,0.0023756688,0.0026007623,0.0034445797,0.003074747,0.0004216599],"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":[0.00046652567,0.00027174494,0.029201524,0.00068230956,0.002844566,0.0006940508,0.00077874766,0.10115982,0.003153409,0.44431436,0.0031710102,0.41326192],"study_design_scores_gemma":[0.0002254992,0.00034549198,0.0074549764,0.00013531868,0.0005596119,0.0005537303,0.0001822465,0.24886346,0.0018392627,0.7360252,0.0037296521,0.000085456726],"about_ca_topic_score_codex":0.0029658843,"about_ca_topic_score_gemma":0.002995293,"teacher_disagreement_score":0.026864346,"about_ca_system_score_codex":0.0008499694,"about_ca_system_score_gemma":0.003438689,"threshold_uncertainty_score":0.14207399},"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":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.007212656,0.00033882144,0.98955756,0.00066830777,0.000073570205,0.00013580837,0.0006887508,0.00052768394,0.00079686596],"genre_scores_gemma":[0.50839305,0.0010663266,0.4727808,0.00073359616,0.0004926766,0.0025336174,0.0030453044,0.0003529125,0.010601695],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9889114,0.007848736,0.00034939818,0.001545372,0.000887171,0.00045785698],"domain_scores_gemma":[0.94421124,0.04867715,0.001556559,0.0032432596,0.0017134547,0.0005982963],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.018793862,0.0015900075,0.0032795176,0.0032501284,0.0014972641,0.002594464,0.0052986694,0.0031058944,0.019158969],"category_scores_gemma":[0.06744993,0.0010576985,0.0036022542,0.004648992,0.0018575364,0.005136489,0.002935976,0.0049861716,0.003115226],"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":[0.0013936636,0.00074279966,0.01970494,0.0009230698,0.0024900339,0.0016417853,0.0015316862,0.12619002,0.0024563554,0.60899323,0.014882632,0.21904978],"study_design_scores_gemma":[0.0001928361,0.00014949501,0.0016191892,0.000070290844,0.0004270434,0.00025485046,0.00011430338,0.6302983,0.00032795893,0.36356628,0.0029268586,0.000052625528],"about_ca_topic_score_codex":0.005304788,"about_ca_topic_score_gemma":0.0049421433,"teacher_disagreement_score":0.019158969,"about_ca_system_score_codex":0.0010295729,"about_ca_system_score_gemma":0.003682928,"threshold_uncertainty_score":0.09939265},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00096066704,0.00024907422,0.9980988,0.0002080206,0.000013928872,0.000020120915,0.000052866497,0.00022842563,0.00016814194],"genre_scores_gemma":[0.12586503,0.0015464177,0.86453205,0.00064977875,0.0002576901,0.0009001021,0.0007727783,0.001264852,0.0042113797],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9925088,0.005195409,0.00032256488,0.0007848365,0.00092441513,0.00026401778],"domain_scores_gemma":[0.95828134,0.03440669,0.0015183243,0.0031700502,0.0021154599,0.00050822366],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.021248562,0.0012518711,0.0029646761,0.0016619461,0.0008000905,0.0021407532,0.0035032805,0.002507447,0.004778947],"category_scores_gemma":[0.09253625,0.0015803913,0.001845041,0.002367911,0.0039634746,0.003114385,0.005289986,0.00497851,0.0013358067],"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":[0.00037927186,0.00011714409,0.0018581266,0.00077170285,0.0003846663,0.00028122976,0.00040012665,0.2663966,0.0036915396,0.5660329,0.008177937,0.15150869],"study_design_scores_gemma":[0.000048033005,0.000040393646,0.00042005588,0.00005754464,0.00002904488,0.00009565257,0.00001931753,0.6569681,0.0007326223,0.33960977,0.0019544738,0.000024935407],"about_ca_topic_score_codex":0.006068021,"about_ca_topic_score_gemma":0.0043676686,"teacher_disagreement_score":0.021248562,"about_ca_system_score_codex":0.0017888525,"about_ca_system_score_gemma":0.0044787517,"threshold_uncertainty_score":0.112374485},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.016790997,0.0006736894,0.9808147,0.00066360267,0.00007402186,0.000071947325,0.0001919265,0.00020134528,0.000517815],"genre_scores_gemma":[0.7440591,0.0027991736,0.23504716,0.00059024745,0.0008846884,0.0010943338,0.0023217862,0.00040195743,0.012801571],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9891944,0.0065144612,0.00046948245,0.0020490377,0.0011638767,0.00060872815],"domain_scores_gemma":[0.88425416,0.1045072,0.005425928,0.0030031195,0.0018153718,0.0009941418],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.026168764,0.0018671063,0.003754021,0.0024519584,0.0007106897,0.003804232,0.0041001374,0.0026061523,0.00510744],"category_scores_gemma":[0.09354675,0.0016706199,0.0033288272,0.0018157795,0.0022676014,0.004449208,0.00275972,0.0045435447,0.00077725376],"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":[0.000475691,0.0001946881,0.005001654,0.00040191467,0.0007677505,0.0003911804,0.00041700134,0.5750296,0.0011637176,0.3641094,0.002423731,0.049623635],"study_design_scores_gemma":[0.000028800037,0.000049080627,0.00037686608,0.000022938957,0.000060884497,0.000053214262,0.000015764495,0.92767256,0.00016547756,0.07106016,0.00047728923,0.000016973301],"about_ca_topic_score_codex":0.0038069328,"about_ca_topic_score_gemma":0.002548344,"teacher_disagreement_score":0.026168764,"about_ca_system_score_codex":0.0016701277,"about_ca_system_score_gemma":0.0026343125,"threshold_uncertainty_score":0.13839531},"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":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.41384083,0.00915489,0.54910415,0.003708093,0.00079994544,0.0005552605,0.004322994,0.013237987,0.005275945],"genre_scores_gemma":[0.7037844,0.0019347001,0.27710867,0.0007240459,0.0002824113,0.0004040591,0.01161659,0.0015761585,0.0025690396],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9870012,0.008321546,0.00084591866,0.0012863153,0.0018885487,0.00065642904],"domain_scores_gemma":[0.90904695,0.07557767,0.0017171907,0.0057003726,0.007044118,0.000913651],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.032615718,0.0018238532,0.002066107,0.0023906275,0.0010466658,0.0021719022,0.0021047406,0.00149988,0.0023992164],"category_scores_gemma":[0.064374544,0.0005069371,0.00199192,0.0029486346,0.0007231671,0.0032986575,0.001608654,0.0020982367,0.00088853994],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0028005526,0.0011806582,0.02699489,0.00066553545,0.0015378932,0.00026436828,0.00019462632,0.6653144,0.0038102649,0.009368393,0.016200364,0.271668],"study_design_scores_gemma":[0.00009728586,0.0006875852,0.002835598,0.000034824006,0.0001344445,0.00013201786,0.00010303195,0.9898684,0.0016887242,0.0030616606,0.0013187134,0.00003760198],"about_ca_topic_score_codex":0.01449851,"about_ca_topic_score_gemma":0.01204405,"teacher_disagreement_score":0.9673843,"about_ca_system_score_codex":0.0019212553,"about_ca_system_score_gemma":0.005429731,"threshold_uncertainty_score":0.17249048},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.034481227,0.00023180948,0.9635734,0.00021333528,0.000045782403,0.000042793465,0.0003159477,0.00029910245,0.00079646864],"genre_scores_gemma":[0.665125,0.0012332201,0.31445816,0.0003433006,0.00022290634,0.0007295469,0.003997804,0.0002882281,0.013601877],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9970571,0.0018389709,0.00013002733,0.00043544918,0.00032954555,0.00020886029],"domain_scores_gemma":[0.97520655,0.020615509,0.00070118293,0.0025339415,0.00066114357,0.00028165404],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.010179839,0.0007720661,0.0027951873,0.0025577182,0.0005784944,0.0020799183,0.0033649157,0.0016150645,0.005686991],"category_scores_gemma":[0.043040305,0.0011041099,0.002911028,0.0023562787,0.0011490077,0.0025622647,0.0028317338,0.002700683,0.0012496815],"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":[0.00087200664,0.0002578146,0.022629406,0.00024474395,0.0010629559,0.0004406412,0.0006572341,0.595694,0.0012139558,0.15401483,0.0042180317,0.21869437],"study_design_scores_gemma":[0.00002580253,0.00006206546,0.00180025,0.000046158337,0.0000718487,0.00012397481,0.000055829914,0.9374467,0.0004077534,0.058617588,0.0013201726,0.000021772548],"about_ca_topic_score_codex":0.0048290985,"about_ca_topic_score_gemma":0.00525644,"teacher_disagreement_score":0.010179839,"about_ca_system_score_codex":0.0005488976,"about_ca_system_score_gemma":0.0021038314,"threshold_uncertainty_score":0.053836763},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2726345,0.0006212168,0.7227499,0.0012455086,0.000116616466,0.000112746064,0.0013836239,0.00037419764,0.0007616924],"genre_scores_gemma":[0.9509454,0.00044911524,0.041238494,0.00023451897,0.00017564073,0.00025875217,0.0018451913,0.00008274256,0.0047701118],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99712664,0.0015167845,0.00016141906,0.00071574614,0.00015067866,0.00032868775],"domain_scores_gemma":[0.9283677,0.06507076,0.002621093,0.0023256924,0.00086184783,0.0007528032],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0130082965,0.0011839647,0.0028970013,0.0010230225,0.000852379,0.0019289578,0.0032459975,0.0026468842,0.0042420304],"category_scores_gemma":[0.03673852,0.0022720906,0.0035688384,0.00096944656,0.0023375421,0.0029335986,0.0022109319,0.0031832173,0.00056216924],"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":[0.0009870685,0.00016651003,0.014977865,0.00015884027,0.0005335734,0.0003440498,0.00041734095,0.9093551,0.00073053845,0.06125717,0.00096272986,0.010109255],"study_design_scores_gemma":[0.000053244417,0.000042612355,0.0015897234,0.000022719632,0.00008979406,0.000037600286,0.000031487452,0.96819085,0.00022307356,0.02950935,0.00017999992,0.000029548175],"about_ca_topic_score_codex":0.01978485,"about_ca_topic_score_gemma":0.0153038185,"teacher_disagreement_score":0.01978485,"about_ca_system_score_codex":0.0013306121,"about_ca_system_score_gemma":0.0024564608,"threshold_uncertainty_score":0.068795264},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.012427414,0.0015748793,0.9846335,0.00028737015,0.00007687077,0.00006164011,0.000115260846,0.00026998812,0.0005530162],"genre_scores_gemma":[0.18895444,0.0038008608,0.8019121,0.00032852124,0.00023159235,0.00045440637,0.0013312364,0.00066919584,0.0023175885],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9881225,0.0092731975,0.0005612442,0.0008101602,0.0010028252,0.00023001256],"domain_scores_gemma":[0.8736209,0.113913596,0.0018493308,0.0056229355,0.004365374,0.00062787184],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0389088,0.0014488146,0.0021430464,0.0025859813,0.0007233432,0.0024391725,0.0039441194,0.0026249292,0.0032803847],"category_scores_gemma":[0.12524697,0.00093249476,0.0025711579,0.0023189236,0.0011580824,0.0041368585,0.0032905985,0.0036819219,0.0008891523],"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":[0.0015879922,0.00026099783,0.012924362,0.0009780794,0.0021470394,0.0002462094,0.0014769678,0.30665275,0.0018803584,0.16261831,0.004762634,0.5044642],"study_design_scores_gemma":[0.00018795436,0.0002460364,0.0043777614,0.00034260642,0.00038802723,0.00027417444,0.0003177467,0.8859771,0.0010772564,0.10300778,0.0036624654,0.00014114125],"about_ca_topic_score_codex":0.006086475,"about_ca_topic_score_gemma":0.005582442,"teacher_disagreement_score":0.0389088,"about_ca_system_score_codex":0.0012720332,"about_ca_system_score_gemma":0.00255356,"threshold_uncertainty_score":0.2057718},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.022750651,0.00029688008,0.9699949,0.00033978492,0.00024074271,0.0013627849,0.0012950208,0.002659705,0.0010595185],"genre_scores_gemma":[0.2290158,0.00016073401,0.75507665,0.0008601877,0.00023000034,0.009227345,0.0022089204,0.0016953574,0.0015250355],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8634897,0.09422692,0.008172097,0.022928765,0.009397023,0.0017855519],"domain_scores_gemma":[0.5044164,0.43108505,0.014357191,0.04326235,0.00525599,0.0016229729],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1275572,0.0020173418,0.0037028189,0.0021848662,0.0023876477,0.00546726,0.0032747223,0.002865011,0.008269959],"category_scores_gemma":[0.25230107,0.0025542597,0.005018299,0.0036914411,0.0048962086,0.0034227544,0.0032023098,0.0043605994,0.0011258706],"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":[0.007258855,0.0017043878,0.11530118,0.007085168,0.017817963,0.0024537647,0.00369622,0.118710704,0.10468846,0.1480489,0.026113244,0.44712114],"study_design_scores_gemma":[0.0019959754,0.002289449,0.051487513,0.0004673419,0.003794073,0.000943611,0.00069308566,0.59272283,0.057769004,0.2447353,0.042668212,0.00043364958],"about_ca_topic_score_codex":0.0022336694,"about_ca_topic_score_gemma":0.004711417,"teacher_disagreement_score":0.1275572,"about_ca_system_score_codex":0.002916076,"about_ca_system_score_gemma":0.008343993,"threshold_uncertainty_score":0.674595},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07517779,0.008236313,0.9054893,0.003906775,0.0012107305,0.00054800045,0.0017748656,0.0013071671,0.0023489736],"genre_scores_gemma":[0.53399986,0.002322632,0.45592934,0.001184907,0.0017499989,0.0008706257,0.001280288,0.0004436851,0.0022187366],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.8225481,0.15860124,0.0046666935,0.007822354,0.0057823826,0.0005792251],"domain_scores_gemma":[0.29526213,0.58471006,0.010251269,0.10168601,0.0065609613,0.0015294643],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.15677772,0.0019321947,0.0047448147,0.0021330249,0.0014233264,0.0028801153,0.006135123,0.0036435123,0.0032289864],"category_scores_gemma":[0.49188703,0.0020641594,0.0063654943,0.0030130553,0.002872274,0.003718213,0.004125524,0.0036713714,0.00049620215],"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":[0.017532496,0.0027302157,0.19257498,0.002851855,0.044026874,0.005168065,0.004150931,0.10507406,0.006627911,0.10732744,0.01226122,0.499674],"study_design_scores_gemma":[0.0053569516,0.0038279472,0.09527752,0.00061073515,0.019145466,0.0026513725,0.00063056056,0.37073228,0.0037528381,0.4803411,0.01713349,0.0005396889],"about_ca_topic_score_codex":0.007544631,"about_ca_topic_score_gemma":0.0062725917,"teacher_disagreement_score":0.84322226,"about_ca_system_score_codex":0.0007582732,"about_ca_system_score_gemma":0.0035910886,"threshold_uncertainty_score":0.82912976},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0030744928,0.009099649,0.984832,0.0013830377,0.00028750813,0.00010521159,0.00033607223,0.0003203863,0.00056150934],"genre_scores_gemma":[0.112437874,0.023313228,0.8507791,0.0011948015,0.001920403,0.0018282194,0.0013653208,0.00043869525,0.006722332],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9699402,0.025641918,0.000855262,0.0018117347,0.001374598,0.0003763178],"domain_scores_gemma":[0.87169296,0.117762715,0.0041558566,0.0034435098,0.0023047456,0.0006403244],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.035014432,0.0037952848,0.006252683,0.004411891,0.0020102484,0.0047839344,0.0070453947,0.006626843,0.004473308],"category_scores_gemma":[0.11081119,0.0026033476,0.0051733307,0.009663556,0.0047270544,0.0046651936,0.004473707,0.008570197,0.0010943803],"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":[0.00024940207,0.00018408251,0.0037499096,0.0012045153,0.001265822,0.00061866693,0.001047141,0.16311397,0.0003818079,0.7268122,0.0075288317,0.093843676],"study_design_scores_gemma":[0.00010826983,0.000101078374,0.0009518224,0.0002931059,0.00034782625,0.00024277868,0.00015621082,0.28676412,0.00014220676,0.70416135,0.0066175447,0.00011365772],"about_ca_topic_score_codex":0.013834915,"about_ca_topic_score_gemma":0.016060019,"teacher_disagreement_score":0.035014432,"about_ca_system_score_codex":0.0024536394,"about_ca_system_score_gemma":0.0042000813,"threshold_uncertainty_score":0.18517625},"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":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008351972,0.00035531033,0.9888694,0.0010279318,0.00008940047,0.00023755222,0.00023374,0.0002440932,0.00059057714],"genre_scores_gemma":[0.4542621,0.0012018265,0.5291574,0.0016430764,0.0007820994,0.00258455,0.0011064049,0.00030408063,0.008958493],"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.96853215,0.021871047,0.001064028,0.0060270317,0.0014861604,0.0010195317],"domain_scores_gemma":[0.8095068,0.16150062,0.007509745,0.015643248,0.004027259,0.0018122563],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.064404935,0.0017646562,0.007844309,0.0020933044,0.0016989415,0.004278684,0.009575137,0.008460655,0.008817641],"category_scores_gemma":[0.17042996,0.0033771887,0.0043444936,0.0030561744,0.0076774,0.008759976,0.0046636835,0.009566694,0.0016846099],"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":[0.00233111,0.0009829008,0.007075313,0.0007856943,0.0013104449,0.00061923714,0.00087077107,0.5127142,0.0016020521,0.32877678,0.004556718,0.13837482],"study_design_scores_gemma":[0.00039081537,0.00021020033,0.00089708134,0.000101648555,0.00020658139,0.00007286052,0.000053572738,0.76836747,0.00034336586,0.22790705,0.0013780227,0.000071311646],"about_ca_topic_score_codex":0.007070441,"about_ca_topic_score_gemma":0.0046808827,"teacher_disagreement_score":0.064404935,"about_ca_system_score_codex":0.0028182208,"about_ca_system_score_gemma":0.0051649497,"threshold_uncertainty_score":0.3406099},"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":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.32559615,0.0038699347,0.6621648,0.0018703057,0.00038186414,0.0002765579,0.0014176695,0.0017999932,0.0026226286],"genre_scores_gemma":[0.8411473,0.0004925599,0.15240286,0.0003548209,0.000411946,0.00035282702,0.0024695839,0.0007928621,0.0015751638],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9125781,0.06544872,0.003567804,0.01130795,0.0057947463,0.0013025465],"domain_scores_gemma":[0.2278013,0.72919875,0.00999632,0.021837857,0.008838653,0.0023270808],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.13442028,0.0020270564,0.0036727027,0.0045847213,0.0021237016,0.004785905,0.0044923597,0.0038276913,0.0039555146],"category_scores_gemma":[0.43342483,0.0017312693,0.0036753176,0.0039824233,0.0059549524,0.0056873737,0.0062166154,0.0040700664,0.0010158708],"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":[0.018689172,0.0006348871,0.31253013,0.0018635241,0.016391868,0.0012701405,0.0018130426,0.24453703,0.021865191,0.036123928,0.008051172,0.33622995],"study_design_scores_gemma":[0.0009873763,0.0022970592,0.061321013,0.00021774422,0.0026201194,0.0010028305,0.0005332849,0.83110195,0.012274645,0.08522505,0.0021646747,0.00025429894],"about_ca_topic_score_codex":0.0022032883,"about_ca_topic_score_gemma":0.0015091333,"teacher_disagreement_score":0.13442028,"about_ca_system_score_codex":0.001307373,"about_ca_system_score_gemma":0.002885395,"threshold_uncertainty_score":0.7108909},"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":"editorial","genre_gemma":"editorial","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"editorial","genre_consensus":"editorial","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0001968753,0.038729116,0.0082456255,0.043682985,0.89475775,0.000061786355,0.00065516384,0.0003932873,0.013277449],"genre_scores_gemma":[0.001049013,0.017422322,0.0020542834,0.016560918,0.91945773,0.0000782411,0.00061305263,0.00053250877,0.04223197],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9975103,0.0004556517,0.00030506056,0.0004779866,0.0010849137,0.00016608174],"domain_scores_gemma":[0.9811218,0.008097616,0.0009342382,0.0010032988,0.0061526094,0.0026904352],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0041294126,0.0025130934,0.0037896622,0.005404288,0.0012481825,0.0056341216,0.0019892638,0.0038865379,0.07748808],"category_scores_gemma":[0.015103934,0.00072226697,0.0024031948,0.0027480808,0.0013969218,0.004389393,0.0022293462,0.008583062,0.049447294],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000016815264,0.000024132416,0.00008081216,0.00021036256,0.000020261117,0.000026346817,0.00000613963,0.00012842294,0.000104806844,0.001336773,0.97602093,0.022024224],"study_design_scores_gemma":[0.000017700702,0.000054376353,0.0008353394,0.00035241252,0.000030111347,0.00016693777,0.000014157282,0.00080732914,0.00010548533,0.009646205,0.98794436,0.000025484565],"about_ca_topic_score_codex":0.0009830788,"about_ca_topic_score_gemma":0.0026600447,"teacher_disagreement_score":0.07748808,"about_ca_system_score_codex":0.0016042114,"about_ca_system_score_gemma":0.0024530164,"threshold_uncertainty_score":0.25922346},"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":"methods","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0016995968,0.0004629825,0.9965004,0.00058275065,0.000070446586,0.000027645381,0.00013072952,0.000150391,0.00037512332],"genre_scores_gemma":[0.3211052,0.0043744347,0.66098636,0.0012480926,0.0018282766,0.00075700064,0.0012442534,0.00040866106,0.008047697],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.97829264,0.014694203,0.0010280565,0.0033399386,0.0020745392,0.00057061116],"domain_scores_gemma":[0.7636452,0.20917687,0.008115535,0.014583673,0.0036533324,0.00082533935],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.037883528,0.0020053927,0.005429219,0.004099748,0.0010491956,0.004082614,0.006561066,0.004396008,0.008224966],"category_scores_gemma":[0.18481405,0.0026292158,0.004812186,0.004923443,0.005441246,0.0077882707,0.0045327223,0.008157437,0.0010382313],"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":[0.00017055088,0.00012280524,0.0017690036,0.0004754996,0.0008876076,0.00032965024,0.0003376217,0.09331886,0.000335903,0.854674,0.0026627402,0.04491571],"study_design_scores_gemma":[0.00005427731,0.000026970492,0.00037905437,0.000081538885,0.00018999736,0.00007912371,0.000027389706,0.20584151,0.0002563073,0.79156786,0.001470472,0.000025500596],"about_ca_topic_score_codex":0.0041524307,"about_ca_topic_score_gemma":0.0020229155,"teacher_disagreement_score":0.9621165,"about_ca_system_score_codex":0.0016234879,"about_ca_system_score_gemma":0.002917801,"threshold_uncertainty_score":0.20034969},"labels":[],"label_agreement":null}]}