{"id":"W4390046685","doi":"10.1093/jamia/ocad226","title":"Semi-supervised ROC analysis for reliable and streamlined evaluation of phenotyping algorithms","year":2023,"lang":"en","type":"article","venue":"Journal of the American Medical Informatics Association","topic":"Sepsis Diagnosis and Treatment","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; University of Toronto","keywords":"Computer science; Receiver operating characteristic; Variance (accounting); Software; Artificial intelligence; Machine learning; Data mining; Algorithm","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":[0.003581796,0.00007444292,0.0004776086,0.0002282518,0.00005909423,0.00001501386,0.00008476042,0.00005369978,0.00002436149],"category_scores_gemma":[0.003203507,0.00004577254,0.000254725,0.0009804247,0.0000377189,0.00009301874,0.00002917291,0.0001255885,0.000001670116],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003338979,"about_ca_system_score_gemma":0.0002621377,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003224147,"about_ca_topic_score_gemma":0.000007234639,"domain_scores_codex":[0.9974558,0.00005776072,0.0007661385,0.00004232188,0.001544068,0.0001338675],"domain_scores_gemma":[0.9966986,0.0004893739,0.001747003,0.0001170848,0.0008376504,0.0001102691],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001502325,0.0003552658,0.8233017,0.0001388812,0.006385356,0.000001349055,0.002676848,0.005560822,0.0001192611,0.00003682665,0.009064962,0.1522085],"study_design_scores_gemma":[0.002376721,0.0002963821,0.3077573,0.0001481245,0.004450689,0.00000434547,0.001771848,0.6823397,0.0003676803,0.0002181444,0.0002172234,0.00005186672],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9941592,0.00007714666,0.00148948,0.00381962,0.0001127332,0.0002463358,0.00001026299,0.000007917918,0.00007728557],"genre_scores_gemma":[0.9962264,0.0004659758,0.002458837,0.0006506159,0.0001117983,0.00001236821,0.00002561274,0.000006770072,0.00004158145],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6767789,"threshold_uncertainty_score":0.3835129,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06715583229088423,"score_gpt":0.3772528395801668,"score_spread":0.3100970072892825,"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."}}