{"meta":{"query_hash":"9d0c7b8de570","filters":{"venue":"Statistical Modelling"},"cohort_total":12,"direct_labels_cover":0,"predictions_cover":12,"exported":12,"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/9d0c7b8de570","api":"https://metacan.xera.ac/api/v1/cohort?venue=Statistical+Modelling"},"results":[{"id":"W2010713129","doi":"10.1177/1471082x0800900203","title":"Clustered binary data with random cluster sizes","year":2009,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Statistical Methods and Bayesian Inference","field":"Mathematics","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Binary data; Random effects model; Cluster (spacecraft); Poisson distribution; Inference; Statistics; Binary number; Mathematics; Best linear unbiased prediction; Moment (physics); Overdispersion; Computer science; Count data; Selection (genetic algorithm); Artificial intelligence","score_opus":0.12350467163338766,"score_gpt":0.38117321083704786,"score_spread":0.2576685392036602,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2010713129","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00046115773,0.000039903694,0.99485743,0.00038539318,0.00009380712,0.00034075862,0.00040728675,0.0001221846,0.0032921007],"genre_scores_gemma":[0.120346494,0.000014544809,0.8789621,0.00033461524,0.00014951712,0.000007500113,0.000076332646,0.000028579554,0.000080315025],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9977783,0.00022537833,0.00049187074,0.0006026303,0.00041525598,0.00048656098],"domain_scores_gemma":[0.9938978,0.004754311,0.0001019403,0.000860151,0.00016161133,0.0002241643],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00063143217,0.00027149514,0.0005120471,0.00005782074,0.00015177306,0.00010821467,0.0004588148,0.00009676539,0.00025867275],"category_scores_gemma":[0.0018408039,0.00019385628,0.00002944885,0.00015309147,0.00015327752,0.00018886865,0.000112035465,0.0003097885,0.00003349166],"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":[0.0009939775,0.0002144723,0.000016701391,0.00010407998,0.000043596872,0.00009095764,0.00012756787,0.0008581101,0.000024089935,0.94664335,0.004962852,0.04592024],"study_design_scores_gemma":[0.0009583831,0.00018815506,0.000026987242,0.000073761556,0.000070487986,0.0000075687485,0.0000183898,0.46876213,0.000009764492,0.5295121,0.0001841982,0.000188064],"about_ca_topic_score_codex":0.000015656828,"about_ca_topic_score_gemma":0.0000050073913,"teacher_disagreement_score":0.467904,"about_ca_system_score_codex":0.00002451837,"about_ca_system_score_gemma":0.00006154618,"threshold_uncertainty_score":0.790523},"labels":[],"label_agreement":null},{"id":"W2041919982","doi":"10.1177/1471082x14566913","title":"The functional linear array model","year":2015,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":54,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"Siemens; Deutsche Forschungsgemeinschaft","keywords":"Covariate; Functional data analysis; Scalar (mathematics); Additive model; Regression analysis; Mathematics; Linear model; Generalized linear model; Linear regression; Proper linear model; Computer science; Applied mathematics; Statistics; Polynomial regression","score_opus":0.3670815074401213,"score_gpt":0.403580841897837,"score_spread":0.03649933445771569,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2041919982","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0004629225,0.000040529623,0.9883381,0.000246207,0.00023235705,0.000114887414,0.000080349004,0.00007176584,0.010412876],"genre_scores_gemma":[0.088875316,0.000009316262,0.9101194,0.00009768375,0.00016271728,0.000022593505,0.000006916043,0.000024844927,0.0006812003],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9984294,0.000117606476,0.00036811322,0.00024870873,0.00047577635,0.0003603643],"domain_scores_gemma":[0.995086,0.004055071,0.000061226965,0.00026650535,0.00025024667,0.0002809311],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009238161,0.00015207737,0.00019619212,0.000020482883,0.00023065424,0.00006109965,0.00015802172,0.00006653286,0.00005938629],"category_scores_gemma":[0.0032696745,0.000098252225,0.000038905677,0.00008503275,0.00018594498,0.000045768244,0.000038169994,0.00028044902,0.00009141629],"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":[0.000048327212,0.000030032283,0.000006709214,0.000009576439,0.000011400834,0.0000023355017,0.00008513713,0.050153054,0.000015928186,0.9429651,0.003485431,0.0031869984],"study_design_scores_gemma":[0.000103973885,0.000023632178,0.000001735658,0.0000058095898,0.000012994957,0.0000017918094,0.00003271674,0.49861872,0.000022842745,0.5006423,0.00046063185,0.00007284263],"about_ca_topic_score_codex":0.000011511068,"about_ca_topic_score_gemma":0.0000026291636,"teacher_disagreement_score":0.44846568,"about_ca_system_score_codex":0.000045448603,"about_ca_system_score_gemma":0.00013871536,"threshold_uncertainty_score":0.40066093},"labels":[],"label_agreement":null},{"id":"W2042186389","doi":"10.1191/1471082x06st121oa","title":"Modelling repeated ordinal reports from multiple informants","year":2006,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Survey Sampling and Estimation Techniques","field":"Mathematics","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Université de Montréal","funders":"Economic and Social Research Council","keywords":"Multivariate statistics; Psychology; Multilevel model; Random effects model; Aggression; Continuation; Ordinal regression; Multivariate analysis; Mathematics; Longitudinal data; Demography; Repeated measures design; Generalized linear model; Statistics; Developmental psychology; Econometrics; Medicine; Computer science; Meta-analysis","score_opus":0.11255771401037017,"score_gpt":0.32497840660770777,"score_spread":0.2124206925973376,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2042186389","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1196241,0.00002570236,0.8775758,0.000013388018,0.0001108145,0.00016816848,0.000117630756,0.0005831733,0.0017812223],"genre_scores_gemma":[0.52480876,0.0000036840759,0.47475922,0.000010447166,0.00007250979,0.000013085547,0.00021941033,0.000025970252,0.00008688525],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9979057,0.000059098864,0.00088677456,0.00038291133,0.00038913844,0.00037638217],"domain_scores_gemma":[0.9975485,0.0015447848,0.00022773138,0.0003993228,0.00017725134,0.00010238254],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006002291,0.00022588483,0.00033442545,0.00009378857,0.0001879756,0.00009110993,0.000104786464,0.00013814261,0.00010225801],"category_scores_gemma":[0.0004179021,0.00021718099,0.000062251405,0.0001330756,0.000059735015,0.00017875084,0.000042343403,0.00024852404,0.000027526188],"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":[0.000059369482,0.00010852336,0.0012588119,0.00005008452,0.000025455367,0.000093212715,0.00014805318,0.952279,0.000057538957,0.042625647,0.00196006,0.0013342928],"study_design_scores_gemma":[0.00010068685,0.000012021377,0.000024013943,0.00004976186,0.000020076255,0.00000827669,0.000010602176,0.573535,0.0003481487,0.42559916,0.00013679467,0.00015545178],"about_ca_topic_score_codex":0.0043296088,"about_ca_topic_score_gemma":0.00002975733,"teacher_disagreement_score":0.4051847,"about_ca_system_score_codex":0.000057229885,"about_ca_system_score_gemma":0.000039269577,"threshold_uncertainty_score":0.88563836},"labels":[],"label_agreement":null},{"id":"W2144666171","doi":"10.1177/1471082x1001100503","title":"Discrete-time survival trees and forests with time-varying covariates","year":2011,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Financial Distress and Bankruptcy Prediction","field":"Business, Management and Accounting","cited_by":40,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"HEC Montréal","funders":"","keywords":"Covariate; Statistics; Discrete time and continuous time; Econometrics; Survival analysis; Bankruptcy; Bankruptcy prediction; Tree (set theory); Mathematics; Accelerated failure time model; Computer science; Economics; Finance","score_opus":0.024052791189226055,"score_gpt":0.21120000310632303,"score_spread":0.187147211917097,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2144666171","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.093018554,0.00003276982,0.8913102,0.00003468934,0.00008277275,0.00013995581,0.00004471945,0.00012705708,0.015209254],"genre_scores_gemma":[0.98686767,0.0000048582015,0.012571782,0.00005038149,0.00025683554,0.000009762049,0.00009802028,0.000027290638,0.000113430375],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99908155,0.000007921541,0.0001855977,0.0002776063,0.00018210568,0.00026523255],"domain_scores_gemma":[0.9996695,0.0000645781,0.00007374143,0.00010501757,0.00006556713,0.000021584408],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00012819804,0.00016225482,0.00019841714,0.000065061686,0.00019108049,0.00014846992,0.00008083833,0.000050773575,0.00032781213],"category_scores_gemma":[0.00003327236,0.00012694942,0.000020747717,0.00011440345,0.00010118168,0.00051547564,0.00006254135,0.000095864016,0.0001310354],"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":[0.0009922948,0.00020081621,0.033963114,0.00030513995,0.000097848715,0.00010397915,0.0002751989,0.011720577,0.00019623592,0.94060457,0.00052922696,0.011010989],"study_design_scores_gemma":[0.00039970272,0.000037066868,0.022920351,0.00007840848,0.00008954355,0.0000013413784,0.000015987456,0.9305397,0.000010660947,0.04548722,0.00018563612,0.00023435701],"about_ca_topic_score_codex":0.0009785035,"about_ca_topic_score_gemma":0.000083534986,"teacher_disagreement_score":0.9188191,"about_ca_system_score_codex":0.000008275944,"about_ca_system_score_gemma":0.0000094023,"threshold_uncertainty_score":0.51768476},"labels":[],"label_agreement":null},{"id":"W2166211535","doi":"10.1177/1471082x0700700406","title":"Worm plot to diagnose fit in quantile regression","year":2007,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":29,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"AGE-WELL","keywords":"Quantile regression; Plot (graphics); Overfitting; Quantile; Covariate; Statistics; Econometrics; Regression; Regression analysis; Mathematics; Parametric statistics; Computer science; Artificial intelligence","score_opus":0.2188968809883266,"score_gpt":0.4481788269679603,"score_spread":0.2292819459796337,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2166211535","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.05831353,0.000022864378,0.9377562,0.000077126715,0.00013319589,0.00024638677,0.000060848015,0.000063872496,0.0033259836],"genre_scores_gemma":[0.4235641,0.000006887273,0.5761721,0.00010179176,0.000047694615,0.000011983215,0.0000025430504,0.000021241964,0.000071671275],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.997974,0.00009336355,0.00059657707,0.00038165593,0.0003774341,0.00057701784],"domain_scores_gemma":[0.99190706,0.0073702494,0.0000608652,0.00026715198,0.000079755104,0.0003149465],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001278233,0.00019495269,0.00035798352,0.00014178049,0.000070429625,0.000037110913,0.00016607018,0.000102832084,0.00036258233],"category_scores_gemma":[0.0033097093,0.00016095482,0.00003271533,0.0002931675,0.000060310023,0.000049481154,0.00007592131,0.00026951288,0.000071168346],"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":[0.00011732531,0.00019001463,0.00035480203,0.00006433781,0.000004683412,0.00012424465,0.00046092,0.00076850323,0.00011905663,0.92725337,0.0008929774,0.06964977],"study_design_scores_gemma":[0.00027615865,0.0001181098,0.0004278819,0.0002604483,0.000011794674,0.000002490526,0.00011248,0.19873445,0.0006152043,0.79884064,0.00035558693,0.0002447767],"about_ca_topic_score_codex":0.00012625812,"about_ca_topic_score_gemma":0.000108495435,"teacher_disagreement_score":0.36525056,"about_ca_system_score_codex":0.00007039755,"about_ca_system_score_gemma":0.000026605609,"threshold_uncertainty_score":0.65635467},"labels":[],"label_agreement":null},{"id":"W2167286520","doi":"10.1177/1471082x0700700204","title":"A measure of partial association for generalized estimating equations","year":2007,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Optimal Experimental Design Methods","field":"Decision Sciences","cited_by":20,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"St. Stephen's University; St. Michael's Hospital","funders":"National Institutes of Health","keywords":"Mathematics; Generalized estimating equation; Statistics; Measure (data warehouse); Estimating equations; Linear regression; Covariate; Regression analysis; Partial correlation; Ordinary least squares; Outcome (game theory); Applied mathematics; Correlation; Estimator","score_opus":0.3184001615592276,"score_gpt":0.5005314592625004,"score_spread":0.1821312977032728,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2167286520","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0024713161,0.000037846057,0.99528134,0.00005766098,0.00033854417,0.00029166634,0.00009994711,0.000024945288,0.001396741],"genre_scores_gemma":[0.4096393,2.2466375e-7,0.5901337,0.000021182139,0.00007963147,0.0000115149005,0.000007319363,0.000008456437,0.00009867],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9969178,0.00023490605,0.0009528924,0.00032338616,0.0012417,0.0003293039],"domain_scores_gemma":[0.9844333,0.014310517,0.00036187406,0.0001957919,0.0005795184,0.00011903007],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.010461778,0.00010642024,0.00030254127,0.000119857825,0.00014628068,0.00008213563,0.00020655553,0.00009060969,0.00011953582],"category_scores_gemma":[0.017727658,0.000092079754,0.00009190915,0.0003146858,0.000049406073,0.00014508658,0.00003168338,0.000093612674,0.000020460675],"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":[0.00020346123,0.00008421636,0.00019174976,0.000011413054,0.000028114111,0.0000014898152,0.00062153686,0.4657313,0.011793467,0.49197343,0.00035123923,0.029008573],"study_design_scores_gemma":[0.00032411417,0.00006025087,0.000018866771,0.000008437252,0.000023838582,2.6529483e-7,0.000096464355,0.81146795,0.0061169243,0.18171713,0.000082778635,0.00008300427],"about_ca_topic_score_codex":0.000035888035,"about_ca_topic_score_gemma":0.000002851628,"teacher_disagreement_score":0.40716797,"about_ca_system_score_codex":0.00011066752,"about_ca_system_score_gemma":0.00006016572,"threshold_uncertainty_score":0.99054646},"labels":[],"label_agreement":null},{"id":"W2286729808","doi":"10.1177/1471082x1001100304","title":"Variable selection in additive models by non-negative garrote","year":2011,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Marine and fisheries research","field":"Environmental Science","cited_by":23,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Selection (genetic algorithm); Feature selection; Variable (mathematics); Computer science; Model selection; Mathematical optimization; Parametric statistics; Series (stratigraphy); Econometrics; Machine learning; Mathematics; Statistics","score_opus":0.03270197921753465,"score_gpt":0.24048484824284386,"score_spread":0.2077828690253092,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2286729808","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0013138468,6.075653e-7,0.6760688,0.000005597988,0.000010310226,0.00012678727,0.000053584798,0.00001129454,0.32240918],"genre_scores_gemma":[0.6357531,0.000011757243,0.36239415,0.00006007046,0.000014581368,0.00005750284,0.000024277446,0.000017739509,0.001666819],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99885,0.000054277112,0.00017713875,0.0003109125,0.00024363498,0.00036407798],"domain_scores_gemma":[0.9996502,0.000122062236,0.000025255898,0.00007399104,0.00001360325,0.000114892406],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00018250881,0.000106057385,0.00012550237,0.000026878392,0.00007779584,0.000018412295,0.00011554359,0.000062306586,0.014824991],"category_scores_gemma":[0.000038437454,0.00010335393,0.00001328481,0.00023737688,0.00013563554,0.0002725124,0.0001002099,0.0002546104,0.00012932664],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0015002089,0.0015996255,0.013829213,0.00009064993,0.000099114244,0.00013239293,0.013248207,0.543348,0.003003391,0.2857382,0.030584361,0.10682664],"study_design_scores_gemma":[0.00014881339,0.00006912045,0.0001643314,0.0000034871798,0.000002776115,7.694324e-7,0.000051643987,0.8099225,0.0002866732,0.1888188,0.00042240423,0.00010866234],"about_ca_topic_score_codex":0.009115142,"about_ca_topic_score_gemma":0.00012918167,"teacher_disagreement_score":0.63443923,"about_ca_system_score_codex":0.00012520251,"about_ca_system_score_gemma":0.00001576383,"threshold_uncertainty_score":0.99748325},"labels":[],"label_agreement":null},{"id":"W2724598028","doi":"10.1177/1471082x17705993","title":"Estimation of partly linear additive hazards model with left-truncated and right-censored data","year":2017,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Statistical Methods and Inference","field":"Mathematics","cited_by":11,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Calgary","funders":"","keywords":"Mathematics; Covariate; Estimator; Spline (mechanical); Linear model; Semiparametric model; Applied mathematics; Parametric statistics; Nonlinear system; Additive model; Semiparametric regression; Inference; Statistics; Econometrics; Computer science; Artificial intelligence","score_opus":0.14440374555615648,"score_gpt":0.40027549152921776,"score_spread":0.2558717459730613,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2724598028","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.010901357,0.00001345768,0.9843978,0.00006539755,0.000030232859,0.00021301059,0.0030996753,0.000034226185,0.0012447871],"genre_scores_gemma":[0.4545646,0.000013479354,0.5452591,0.000008292718,0.000019226916,0.00000376687,0.0000813621,0.000017210246,0.000032952397],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99841905,0.00007855585,0.0004253569,0.00045555757,0.0003444236,0.00027706832],"domain_scores_gemma":[0.9967862,0.0016144519,0.00027408884,0.00096178544,0.00020336873,0.00016011663],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000459998,0.00020509765,0.0004390853,0.000030244648,0.00028050444,0.00008384587,0.00037454464,0.00008945315,0.000088867426],"category_scores_gemma":[0.0024398116,0.00015543295,0.000015899543,0.000029391393,0.00049770455,0.000228436,0.00018587093,0.00021017919,0.0000040627338],"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":[0.00014888505,0.00011428713,0.000047274436,0.00018165761,0.000068968286,0.000015925512,0.00021162404,0.039598532,0.00003132207,0.93613714,0.00029279038,0.023151614],"study_design_scores_gemma":[0.0002643669,0.0000633405,0.00005010211,0.00009655018,0.00009886213,0.0000029544167,0.000015251302,0.5948401,0.00023953203,0.40419942,0.000004286006,0.00012521064],"about_ca_topic_score_codex":0.00007036303,"about_ca_topic_score_gemma":0.00001589938,"teacher_disagreement_score":0.5552416,"about_ca_system_score_codex":0.000013215659,"about_ca_system_score_gemma":0.00011177452,"threshold_uncertainty_score":0.63383716},"labels":[],"label_agreement":null},{"id":"W2782124212","doi":"10.1177/1471082x17746538","title":"Frailty modelling for multitype recurrent events in clinical trials","year":2018,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Health Systems, Economic Evaluations, Quality of Life","field":"Economics, Econometrics and Finance","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian VIGOUR Centre; University of Alberta","funders":"","keywords":"Event (particle physics); Computer science; Feature (linguistics); Event data; Data science; Medicine; Machine learning","score_opus":0.8667122139808082,"score_gpt":0.6079313804695468,"score_spread":0.25878083351126147,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2782124212","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.025452584,0.00050665694,0.96764255,0.0019753966,0.0015203828,0.0012481555,0.0009975877,0.000038826485,0.0006178369],"genre_scores_gemma":[0.667869,0.00015792942,0.32838562,0.0016900207,0.0013799354,0.00017497854,0.0001119885,0.00005242698,0.00017816681],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9888324,0.0010413122,0.008567831,0.0008503278,0.00011352545,0.00059456594],"domain_scores_gemma":[0.98703337,0.010370805,0.0017502694,0.00041448243,0.00014270502,0.00028835947],"candidate_categories":["metaresearch","metaepi_narrow","insufficient_payload"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.061518148,0.00022156375,0.0017086394,0.0002342147,0.00018051654,0.000040872223,0.0002501149,0.0002561694,0.0003508681],"category_scores_gemma":[0.017140122,0.0002664743,0.00021104308,0.00013979197,0.000115233226,0.00023988738,0.00004457126,0.00031848965,0.0012997498],"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":[0.00042036883,0.00047106334,0.0078393705,0.00041145665,0.000114458046,0.0000010723944,0.0011970152,0.12684812,5.097752e-7,0.8513751,0.0064740623,0.004847401],"study_design_scores_gemma":[0.00105171,0.00014363133,0.00031036997,0.00008062853,0.00000849202,3.0219087e-7,0.000056700002,0.7460998,8.275898e-7,0.24411747,0.0079122465,0.00021778836],"about_ca_topic_score_codex":0.00049212476,"about_ca_topic_score_gemma":0.00011978728,"teacher_disagreement_score":0.64241636,"about_ca_system_score_codex":0.0002935044,"about_ca_system_score_gemma":0.00014786216,"threshold_uncertainty_score":0.9999787},"labels":[],"label_agreement":null},{"id":"W2905530915","doi":"10.1177/1471082x18810114","title":"Component-based regularization of a multivariate GLM with a thematic partitioning of the explanatory variables","year":2018,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Remote Sensing in Agriculture","field":"Environmental Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"European Commission; Biodiversa+; Canadian Institute for Advanced Research","keywords":"Mathematics; Generalized linear model; Linear model; Regularization (linguistics); Linear regression; Design matrix; Contrast (vision); Statistics; Covariate; Applied mathematics; Computer science; Artificial intelligence","score_opus":0.01355865584465346,"score_gpt":0.21337092113268516,"score_spread":0.1998122652880317,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2905530915","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12291385,0.0000023375771,0.8755737,0.000033648845,0.000030068612,0.0001371955,0.000012287099,0.000011217073,0.0012856756],"genre_scores_gemma":[0.76842,2.578272e-7,0.23151244,0.000021381726,0.000009755045,9.240235e-7,0.000006656887,0.000006550476,0.000022078015],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990921,0.0000887606,0.00022058596,0.00015594592,0.00031364668,0.00012898068],"domain_scores_gemma":[0.9994121,0.00017859506,0.00015379276,0.00018744176,0.0000338451,0.000034178473],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00015792792,0.000084163636,0.00013194587,0.000011186838,0.00009885077,0.000008851871,0.00010778284,0.000038277438,0.00010793726],"category_scores_gemma":[0.0000570313,0.000048041467,0.000019729878,0.00016694743,0.0004306189,0.00004636416,0.000034288252,0.00006882712,0.000006810106],"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":[0.000052997017,0.00009265515,0.00096043566,0.000057925343,0.000017648368,0.0000012604194,0.00069046573,0.94450235,0.026684841,0.026557494,0.000073569136,0.00030832915],"study_design_scores_gemma":[0.00020558784,0.00005184482,0.0018882102,0.00019833515,0.00003165821,0.0000019328459,0.000042898893,0.97548354,0.012383831,0.0096242875,0.000015476206,0.00007240703],"about_ca_topic_score_codex":0.00020465166,"about_ca_topic_score_gemma":0.000025795242,"teacher_disagreement_score":0.6455061,"about_ca_system_score_codex":0.000035357512,"about_ca_system_score_gemma":0.000013379807,"threshold_uncertainty_score":0.19590741},"labels":[],"label_agreement":null},{"id":"W3165324759","doi":"10.1177/1471082x211008011","title":"Bayesian adjustment for measurement error in an offset variable in a Poisson regression model","year":2021,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Traffic and Road Safety","field":"Engineering","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Statistics; Observational error; Poisson regression; Markov chain Monte Carlo; Crash; Bayesian probability; Econometrics; Errors-in-variables models; Regression analysis; Poisson distribution; Population; Random effects model; Computer science; Mathematics; Demography; Medicine","score_opus":0.05915737857479945,"score_gpt":0.2796012220572382,"score_spread":0.22044384348243876,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3165324759","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004592083,0.00034557833,0.9940993,0.00004782117,0.00013276869,0.00022266786,0.000110730994,0.000072922514,0.00037612618],"genre_scores_gemma":[0.6995751,0.0000322273,0.30018738,0.000020477204,0.000026470465,0.000045319677,0.0000721448,0.000025040392,0.00001582585],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987641,0.000036585257,0.00033370245,0.0002770858,0.00023583269,0.00035269107],"domain_scores_gemma":[0.99958575,0.00005912351,0.000015137221,0.00016412561,0.0000627921,0.00011304776],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00028890828,0.00015020857,0.00022758101,0.00006462033,0.000035571033,0.000015771475,0.00006809472,0.00009803372,0.000016235983],"category_scores_gemma":[0.000023454091,0.00014379082,0.00002404125,0.00012881498,0.000011065437,0.000088646615,0.000015169401,0.00018225914,0.0000016918958],"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":[0.000047031404,0.00010601287,0.000029046081,0.00011540969,0.000007620179,0.000012646627,0.00029168825,0.97130585,0.00033541548,0.023548359,0.00011491152,0.0040860134],"study_design_scores_gemma":[0.0006734233,0.000026889973,0.000115401024,0.0001867985,0.0000144342,0.0000011349049,0.00008850795,0.98108995,0.0001179675,0.017446415,0.00006484253,0.00017422273],"about_ca_topic_score_codex":0.000026094916,"about_ca_topic_score_gemma":0.00024327355,"teacher_disagreement_score":0.694983,"about_ca_system_score_codex":0.00023056098,"about_ca_system_score_gemma":0.00008262858,"threshold_uncertainty_score":0.58636194},"labels":[],"label_agreement":null},{"id":"W3217739957","doi":"10.1177/1471082x211059233","title":"Bayesian analysis of two-part nonlinear latent variable model: Semiparametric method","year":2021,"lang":"en","type":"article","venue":"Statistical Modelling","topic":"Bayesian Methods and Mixture Models","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Manitoba","funders":"","keywords":"Covariate; Mathematics; Collinearity; Econometrics; Statistics; Semiparametric regression; Latent variable; Bayesian probability; Semiparametric model; Population; Parametric statistics","score_opus":0.03581909957611684,"score_gpt":0.3270863331790412,"score_spread":0.29126723360292434,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3217739957","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00005537468,0.00021208302,0.9968618,0.00010293488,0.0001416238,0.00009628862,0.00015982387,0.00007876286,0.0022913173],"genre_scores_gemma":[0.017379662,0.00005150145,0.98192686,0.00022481765,0.000042868644,0.000009190861,0.000052316056,0.000021388347,0.00029141444],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9970157,0.00039089174,0.0007114499,0.0008411138,0.00053056265,0.0005102377],"domain_scores_gemma":[0.99724734,0.0010321593,0.0001600743,0.0008988025,0.00038279337,0.00027883882],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011994769,0.00024439205,0.0007785851,0.00043103076,0.000104134364,0.00012062866,0.00052203395,0.00012273387,0.00008974204],"category_scores_gemma":[0.00018236352,0.00023434356,0.00021470363,0.0036132196,0.00004736976,0.00021264204,0.00023720929,0.0003063276,0.0000047645804],"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":[0.000003440375,0.0000672859,0.000009391432,0.000018757913,0.00019958464,0.000017791794,0.00006689982,0.55500686,0.00015633866,0.43650472,0.00003125149,0.007917698],"study_design_scores_gemma":[0.00016497585,0.0000184061,0.0000015167963,0.000014059251,0.0004899316,0.0000046942882,0.0000017833007,0.72843236,0.00073077803,0.26989627,0.000056793182,0.00018841393],"about_ca_topic_score_codex":0.00008409232,"about_ca_topic_score_gemma":0.0000056249723,"teacher_disagreement_score":0.17342553,"about_ca_system_score_codex":0.000046928155,"about_ca_system_score_gemma":0.00025885119,"threshold_uncertainty_score":0.95562536},"labels":[],"label_agreement":null}]}