{"id":"W4413196475","doi":"10.2196/78082","title":"Preprocessing Large-Scale Conversational Datasets: A Framework and Its Application to Behavioral Health Transcripts","year":2025,"lang":"en","type":"article","venue":"JMIR Formative Research","topic":"Mental Health via Writing","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Preprint; Scale (ratio); Preprocessor; Computer science; Computational biology; Natural language processing; Psychology; Artificial intelligence; World Wide Web; Biology; Geography; Cartography","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002027826,0.0001106665,0.0001801932,0.0003336025,0.0006605601,0.00005812391,0.0002194404,0.0001141561,0.0001947013],"category_scores_gemma":[0.00002781368,0.0001154137,0.00002032043,0.0007806915,0.00007343774,0.0002999195,0.0001391153,0.0006278561,0.0002444414],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002885215,"about_ca_system_score_gemma":0.0001939425,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001499238,"about_ca_topic_score_gemma":0.00007240051,"domain_scores_codex":[0.9976577,0.000409838,0.0003900498,0.0004113004,0.0004545943,0.0006764895],"domain_scores_gemma":[0.9989988,0.0002419036,0.00007258365,0.0002858641,0.0001553376,0.0002454744],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.001273224,0.00296699,0.01545372,0.005178498,0.00008095974,0.000008252086,0.5096172,0.000004840782,0.001124315,0.1609452,0.07509567,0.2282512],"study_design_scores_gemma":[0.009369589,0.003755857,0.3918259,0.005618383,0.00003321861,0.00005866779,0.3135758,0.01064944,0.003354983,0.02121378,0.2391847,0.001359734],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9256463,0.001753028,0.04831888,0.0111718,0.0003329707,0.005146296,0.001842454,0.0001349758,0.005653317],"genre_scores_gemma":[0.9958795,0.00001458329,0.0007892789,0.001319717,0.00004679021,0.001351346,0.0003617932,0.00001236568,0.0002246369],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3763721,"threshold_uncertainty_score":0.5080561,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1010176189570137,"score_gpt":0.5400946352524106,"score_spread":0.4390770162953969,"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."}}