{"id":"W4399462990","doi":"10.2196/52896","title":"Unsupervised Feature Selection to Identify Important ICD-10 and ATC Codes for Machine Learning on a Cohort of Patients With Coronary Heart Disease: Retrospective Study","year":2024,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary","funders":"","keywords":"Feature selection; Autoencoder; Artificial intelligence; Computer science; Overfitting; Machine learning; Feature (linguistics); Unsupervised learning; Pattern recognition (psychology); Data mining; Artificial neural network","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[{"model":"gemma","categories":[],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low","status":"direct model label, unvalidated"},{"model":"gpt","categories":[],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low","status":"direct model label, unvalidated"}],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001642017,0.000188966,0.0004117121,0.000203564,0.000496632,0.00002434444,0.0001037387,0.0002145625,0.000321029],"category_scores_gemma":[0.000981668,0.0001282771,0.00003987496,0.0003669042,0.00006189664,0.000244719,0.00007311221,0.001093564,0.0000369992],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001976855,"about_ca_system_score_gemma":0.0005013641,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002496784,"about_ca_topic_score_gemma":0.00004019419,"domain_scores_codex":[0.99708,0.000147252,0.001022898,0.0001620584,0.001203475,0.0003843596],"domain_scores_gemma":[0.998166,0.0004621776,0.0002362493,0.0001517574,0.0003307628,0.0006530075],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001001437,0.0001689511,0.9719295,0.002664007,0.00005869744,0.000001902048,0.01056085,0.000006752818,6.942415e-7,0.0002230682,0.01189475,0.001489426],"study_design_scores_gemma":[0.001759005,0.002434409,0.9578081,0.001419949,0.0000637555,0.000001138402,0.002536946,0.02274321,7.442445e-7,0.00004956,0.01103655,0.000146617],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9928467,0.00005780651,0.0007387613,0.001443319,0.0002131279,0.004186317,0.0001486958,0.0001614526,0.000203859],"genre_scores_gemma":[0.9957488,0.00002458247,0.0004433692,0.002076782,0.0001155618,0.000678326,0.0003589145,0.00002260347,0.0005310375],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02273646,"threshold_uncertainty_score":0.5230989,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04065006443302908,"score_gpt":0.4207781042163397,"score_spread":0.3801280397833106,"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."}}