{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002760207,0.00009064461,0.0003182395,0.0001599706,0.00001783748,0.000002991575,0.00004170792,0.0000356551,0.00009641746],"category_scores_gemma":[0.0006818871,0.00007904268,0.0000241548,0.000494517,0.00003133261,0.0001492833,0.00002105159,0.0002329377,4.483658e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001569613,"about_ca_system_score_gemma":0.0004157048,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.09643147,"about_ca_topic_score_gemma":0.05419184,"domain_scores_codex":[0.9986513,0.00007342136,0.0004502053,0.000172507,0.0005485342,0.0001040241],"domain_scores_gemma":[0.9992645,0.00006962335,0.0001481112,0.0001715088,0.0002668516,0.00007933825],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00009689487,0.0001678293,0.9784329,0.0001792288,0.00002128473,0.00002263314,0.00305763,0.001377433,0.007113742,0.0003925343,0.001157701,0.007980221],"study_design_scores_gemma":[0.003605025,0.000268624,0.9638575,0.001405041,0.0001760126,0.00002374673,0.004435907,0.002086394,0.02187103,0.00007412065,0.002046419,0.0001501511],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9939742,0.0005104934,0.003514281,0.001095829,0.0004052804,0.0002039352,0.0000051104,0.00001121785,0.0002796692],"genre_scores_gemma":[0.9990885,0.0001267479,0.00008490778,0.0001079579,0.00006153858,0.0000170466,0.0003654008,0.00001118992,0.0001367217],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04223963,"threshold_uncertainty_score":0.9630667,"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."}}