{"id":"W2784112886","doi":"10.2196/publichealth.7726","title":"Detecting Novel and Emerging Drug Terms Using Natural Language Processing: A Social Media Corpus Study","year":2018,"lang":"en","type":"article","venue":"JMIR Public Health and Surveillance","topic":"Mental Health via Writing","field":"Psychology","cited_by":32,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Institute on Drug Abuse","keywords":"Social media; Public health; Data science; Drug; Field (mathematics); Internet privacy; Computer science; Sample (material); Natural (archaeology); Ask price; Illicit drug; Psychology; Medicine; Pharmacology; World Wide Web; Pathology; Geography; 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":[0.002143483,0.0004935483,0.0003233742,0.002698344,0.001039849,0.001088765,0.0007551034,0.001125396,0.001005548],"category_scores_gemma":[0.01144676,0.0001963284,0.0006777111,0.00224428,0.001234863,0.002725471,0.001410499,0.001049677,0.0005089112],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008808152,"about_ca_system_score_gemma":0.0005775008,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006270214,"about_ca_topic_score_gemma":0.01254565,"domain_scores_codex":[0.9981592,0.0008765582,0.00021482,0.0003292912,0.0003322314,0.00008792835],"domain_scores_gemma":[0.9877676,0.008685305,0.001272637,0.0009623761,0.001058808,0.0002532819],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002350489,0.005368418,0.4438139,0.003895522,0.0008143124,0.007204604,0.02772376,0.01310134,0.05066234,0.01014872,0.04985703,0.3850595],"study_design_scores_gemma":[0.0004806329,0.001402206,0.589179,0.0005581637,0.0006327791,0.006056326,0.04119295,0.215798,0.02494965,0.008809477,0.110558,0.0003827414],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9847153,0.0005825685,0.005054116,0.001186707,0.0001328925,0.0003364835,0.006148451,0.000150262,0.00169306],"genre_scores_gemma":[0.9523336,0.0006974899,0.0251722,0.0004802531,0.0002499337,0.0006440171,0.0185223,0.00007050874,0.001829612],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.006270214,"threshold_uncertainty_score":0.01246738,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05855809624271563,"score_gpt":0.4035978985095184,"score_spread":0.3450398022668028,"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."}}