{"id":"W2906789718","doi":"10.2196/12414","title":"Data Analysis and Visualization of Newspaper Articles on Thirdhand Smoke: A Topic Modeling Approach","year":2019,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Smoking Behavior and Cessation","field":"Medicine","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Social Science Fund of China","keywords":"Newspaper; Visualization; Data visualization; Computer science; Data science; Topic model; Information retrieval; Data mining; Advertising; Business","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003860298,0.001052014,0.0008276781,0.02313921,0.0006851497,0.003638026,0.0007083285,0.0007716839,0.003317369],"category_scores_gemma":[0.01241061,0.0003678363,0.001607764,0.01656214,0.0003601345,0.001773585,0.001179777,0.0008729717,0.001609655],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007934009,"about_ca_system_score_gemma":0.0008429772,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004550462,"about_ca_topic_score_gemma":0.006751624,"domain_scores_codex":[0.9978483,0.0008203969,0.0003232679,0.0004302512,0.0004460191,0.0001317373],"domain_scores_gemma":[0.9825193,0.0123414,0.001944561,0.0006112556,0.002217933,0.0003656156],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.002106376,0.0009396882,0.2204461,0.008186829,0.00103834,0.002142439,0.02615718,0.01343949,0.02443269,0.005288073,0.06651036,0.6293125],"study_design_scores_gemma":[0.0002321979,0.0008773469,0.4683946,0.001587914,0.001129506,0.001812791,0.03092807,0.3177068,0.0225036,0.01131688,0.1430802,0.0004301684],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6972592,0.005876183,0.1011329,0.003098988,0.0004594977,0.00243628,0.1644112,0.01308301,0.0122428],"genre_scores_gemma":[0.6335516,0.001907118,0.2865308,0.0001455365,0.0004039675,0.003329107,0.06970934,0.0004466031,0.003975902],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02313921,"threshold_uncertainty_score":0.02041543,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06729296493470174,"score_gpt":0.3616715974657648,"score_spread":0.294378632531063,"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."}}