{"id":"W4391055787","doi":"10.1093/jamiaopen/ooae001","title":"A platform for connecting social media data to domain-specific topics using large language models: an application to student mental health","year":2024,"lang":"en","type":"article","venue":"JAMIA Open","topic":"Mental Health via Writing","field":"Psychology","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Health Canada","keywords":"Social media; Computer science; Parsing; Theme (computing); Process (computing); Data science; Domain (mathematical analysis); Key (lock); Similarity (geometry); Human–computer interaction; World Wide Web; Natural language processing; Artificial intelligence; Information retrieval; Programming language","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"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":[0.002762955,0.0009978546,0.0004057735,0.001755633,0.0006816701,0.001585799,0.001410679,0.0008716073,0.005995503],"category_scores_gemma":[0.009219112,0.0004682082,0.001151377,0.0009176422,0.0005298003,0.002444544,0.003017629,0.001372616,0.002706503],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006500673,"about_ca_system_score_gemma":0.001147433,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003732204,"about_ca_topic_score_gemma":0.005729223,"domain_scores_codex":[0.9988878,0.0003440524,0.0001204803,0.0003254397,0.0002762682,0.00004588181],"domain_scores_gemma":[0.9956124,0.002723942,0.0003753424,0.0004992334,0.0004777009,0.0003114142],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003084086,0.002369552,0.03035153,0.001688174,0.0009243026,0.003626606,0.009134012,0.03998194,0.08954437,0.01835923,0.08439159,0.7165446],"study_design_scores_gemma":[0.0003931335,0.0007836417,0.01043156,0.0001896973,0.000189295,0.0007688472,0.001173255,0.8501818,0.02946598,0.03139351,0.07477459,0.0002546441],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0749801,0.000203659,0.7775921,0.00178788,0.000238987,0.001873341,0.007964663,0.1322273,0.003132081],"genre_scores_gemma":[0.1731772,0.0001572435,0.8092398,0.0004971494,0.0001138862,0.001412143,0.007835566,0.00273958,0.004827447],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005995503,"threshold_uncertainty_score":0.02005696,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2664370795296838,"score_gpt":0.5197390405712066,"score_spread":0.2533019610415227,"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."}}