{"id":"W4382362705","doi":"10.21203/rs.3.rs-3085472/v1","title":"Drug and natural health product data collection and curation in the Canadian Longitudinal Study on Aging (CLSA)","year":2023,"lang":"en","type":"preprint","venue":"Research Square","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Dalhousie University; University of Toronto; McMaster University; Université de Sherbrooke; Western University","funders":"","keywords":"Data collection; Sample (material); Gold standard (test); Product (mathematics); Computer science; Drug; Medicine; Statistics; Mathematics; Pharmacology; Chemistry","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00350015,0.0001200841,0.0001424507,0.0002334806,0.0004278645,0.0001720633,0.0003608351,0.00007800615,7.849048e-7],"category_scores_gemma":[0.0008198534,0.00008740654,0.00001256803,0.0002414734,0.0001651775,0.000003910025,0.0007358958,0.000896177,0.000001366452],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009218975,"about_ca_system_score_gemma":0.000760122,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.1526448,"about_ca_topic_score_gemma":0.7931599,"domain_scores_codex":[0.9976619,0.0007296469,0.0001515445,0.0007150199,0.0003901935,0.0003516897],"domain_scores_gemma":[0.9990327,0.00009504845,0.00004009586,0.0006461625,0.00008532212,0.0001006529],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002434569,0.00038219,0.7976716,0.001183332,0.0001773872,0.0001469351,0.009775666,0.00002650808,0.0002416938,0.00007331535,0.110347,0.07973102],"study_design_scores_gemma":[0.000309953,0.0004199617,0.9899625,0.0002151975,0.000004840513,0.00001088546,0.004220097,0.0004188434,0.00003751503,0.0001598367,0.004099045,0.0001413021],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9753687,0.004936406,0.00001085556,0.01800382,0.0002308067,0.001301514,0.00009103669,0.00001550647,0.00004128654],"genre_scores_gemma":[0.9978445,0.0007990479,0.00006273331,0.0000562921,0.0002177397,0.00007617944,0.0006973064,0.0000117528,0.0002344333],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6405151,"threshold_uncertainty_score":0.8529978,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2850080947986918,"score_gpt":0.4962583178454439,"score_spread":0.2112502230467521,"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."}}