{"id":"W3133821461","doi":"10.37349/emed.2021.00033","title":"A machine learning approach to identify correlates of current e-cigarette use in Canada","year":2021,"lang":"en","type":"article","venue":"Exploration of Medicine","topic":"Smoking Behavior and Cessation","field":"Medicine","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Ontario; University of Toronto; Toronto General Hospital; Centre for Addiction and Mental Health; Toronto Public Health; Ontario Tobacco Research Unit","funders":"Canadian Institutes of Health Research","keywords":"Overfitting; Multicollinearity; Logistic regression; Electronic cigarette; Population; Medicine; Demography; Machine learning; Psychology; Regression analysis; Artificial intelligence; Environmental health; Computer science","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003027509,0.0006998068,0.0006420583,0.003984792,0.002180016,0.001271472,0.001326982,0.0006367957,0.002091126],"category_scores_gemma":[0.01172265,0.0003170257,0.001124167,0.003761961,0.0005069132,0.0005100474,0.0009359042,0.001200774,0.0003110787],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01310611,"about_ca_system_score_gemma":0.02517989,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9401281,"about_ca_topic_score_gemma":0.9084912,"domain_scores_codex":[0.9987913,0.0004335746,0.00007876261,0.0001922296,0.0002923702,0.0002117352],"domain_scores_gemma":[0.9957261,0.001764436,0.0003113723,0.0001253883,0.00181697,0.0002558244],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001842711,0.0002249738,0.8963189,0.00009823952,0.0002976941,0.0001641217,0.0004326701,0.02818086,0.000262279,0.00154806,0.003966583,0.06832146],"study_design_scores_gemma":[0.00005047524,0.0001553444,0.3389932,0.0001166876,0.0001753136,0.0001275598,0.001128654,0.653612,0.0003903455,0.002326429,0.002879088,0.0000449023],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.946981,0.001381577,0.03707462,0.002077928,0.00009546693,0.0006357983,0.00476543,0.0003926507,0.006595645],"genre_scores_gemma":[0.9739993,0.0003667441,0.02205289,0.0001110421,0.00001654898,0.0002088207,0.001715297,0.0000168412,0.001512407],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05987191,"threshold_uncertainty_score":0.1204489,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09844205310195694,"score_gpt":0.3359191153290826,"score_spread":0.2374770622271257,"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."}}