{"id":"W4367048213","doi":"10.1007/s10654-023-00997-3","title":"Evaluating the harmonisation potential of diverse cohort datasets","year":2023,"lang":"en","type":"article","venue":"European Journal of Epidemiology","topic":"Health, Environment, Cognitive Aging","field":"Environmental Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Medical Research Foundation; Medical Research Council; UK Research and Innovation; Wellcome Trust; McGill University","keywords":"Comparability; Rigour; Data mining; Computer science; Granularity; Data science; Mathematics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.02875743,0.00009776215,0.0002528084,0.00005485203,0.0001603533,0.000003944632,0.0003873646,0.00002235004,0.0008248929],"category_scores_gemma":[0.005044327,0.00006748177,0.00009061669,0.0001627558,0.0003628191,0.0001486385,0.0003275812,0.0003141341,0.0008307509],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007416801,"about_ca_system_score_gemma":0.00001450931,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004332726,"about_ca_topic_score_gemma":0.000002097364,"domain_scores_codex":[0.9926726,0.005599658,0.0009152673,0.0002301921,0.0002792062,0.0003031036],"domain_scores_gemma":[0.9974382,0.00118106,0.0009761262,0.0002840497,0.00001334226,0.000107223],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001616903,0.00008293343,0.5832421,0.00001884004,0.0001367231,0.0003537157,0.0009907947,0.1144861,0.04558032,0.00008014507,0.04532152,0.2095452],"study_design_scores_gemma":[0.0003552457,0.0003236273,0.9852533,0.0000208951,0.00005666494,0.00007383655,0.0001993586,0.006571689,0.0001094534,0.0003722822,0.006592021,0.00007158807],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9899477,0.00005017466,0.005864508,0.002226007,0.0002919493,0.0001543,0.00002633938,0.000009686329,0.001429323],"genre_scores_gemma":[0.9959158,0.0002712227,0.002426323,0.001131178,0.0001532585,9.370964e-7,0.00003041921,0.00001547152,0.00005535697],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4020112,"threshold_uncertainty_score":0.9999472,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1693394519202947,"score_gpt":0.3948474087422008,"score_spread":0.2255079568219061,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}