{"id":"W4415323123","doi":"10.34105/j.kmel.2025.17.031","title":"From data to insights: Machine learning in thematic analysis of complex health conditions on social media","year":2025,"lang":"en","type":"article","venue":"Knowledge Management & E-Learning An International Journal","topic":"Data-Driven Disease Surveillance","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Thompson Rivers University","funders":"Thompson Rivers University","keywords":"Thematic analysis; Social media; Thematic map; Big data; Filter (signal processing); Process (computing); Class (philosophy); Focus (optics)","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":[0.0008181836,0.0001778671,0.0005757086,0.002217553,0.0001946713,0.0001019037,0.0008817911,0.00003756611,0.0007638347],"category_scores_gemma":[0.0006033832,0.0001701281,0.0001369672,0.001112034,0.00005446256,0.0001985077,0.0005026751,0.0005729336,0.00006675212],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000358621,"about_ca_system_score_gemma":0.0001234504,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008183262,"about_ca_topic_score_gemma":0.0007415878,"domain_scores_codex":[0.9975385,0.0004905126,0.0007366181,0.0004137796,0.0005994844,0.0002211163],"domain_scores_gemma":[0.9985076,0.0003265615,0.0003634435,0.0003807723,0.0002571563,0.0001644721],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.002707401,0.008713676,0.3894078,0.0006280345,0.03644989,0.0008466428,0.04072751,0.04103897,0.0006135913,0.03419449,0.06308757,0.3815844],"study_design_scores_gemma":[0.002443818,0.0001310712,0.877652,0.0009210097,0.000721929,0.000003499284,0.005526586,0.07449451,0.000003937971,0.0003640002,0.03754562,0.0001920386],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8930268,0.0007986079,0.007804875,0.007577973,0.001664517,0.0008953532,0.001489196,0.0003034061,0.08643931],"genre_scores_gemma":[0.9897437,0.0001481476,0.001097752,0.0005416392,0.0002721974,0.000009974568,0.007456449,0.0000184632,0.0007116924],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4882441,"threshold_uncertainty_score":0.8363454,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0916651111064284,"score_gpt":0.4145368601336094,"score_spread":0.322871749027181,"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."}}