{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001832981,0.0001558197,0.0002730478,0.00008537386,0.0005362383,0.0002817572,0.0009846733,0.00008412456,0.0001213173],"category_scores_gemma":[0.000009944482,0.0001666091,0.0000286407,0.0002360237,0.000008583506,0.0004326527,0.0009207787,0.0001732788,0.00009910495],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006108924,"about_ca_system_score_gemma":0.00009416858,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006894888,"about_ca_topic_score_gemma":0.002789627,"domain_scores_codex":[0.997905,0.00008628102,0.000462285,0.0007443306,0.0002355833,0.0005665073],"domain_scores_gemma":[0.998911,0.0001271615,0.00008966877,0.0005966505,0.0000243667,0.0002511356],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003375337,0.0003732841,0.00007526329,0.0002551309,0.00006887745,0.00001646544,0.3893769,0.00002595163,0.001261944,0.1788173,0.02416064,0.4052307],"study_design_scores_gemma":[0.002762843,0.0003980497,0.001002782,0.0003046588,0.00001988724,0.00003940128,0.3229067,0.01578518,0.00007018147,0.00179408,0.6543125,0.0006037377],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8874962,0.0007451719,0.09041795,0.006310176,0.001809701,0.007503832,0.002236837,0.0001853226,0.003294827],"genre_scores_gemma":[0.9138947,0.000008163688,0.07417082,0.005623329,0.002822867,0.0007133847,0.002436148,0.000128705,0.0002018655],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6301519,"threshold_uncertainty_score":0.6794121,"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."}}