{"id":"W4385264305","doi":"10.2196/47736","title":"Comparing Explainable Machine Learning Approaches With Traditional Statistical Methods for Evaluating Stroke Risk Models: Retrospective Cohort Study","year":2023,"lang":"en","type":"article","venue":"JMIR Cardio","topic":"Acute Ischemic Stroke Management","field":"Medicine","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Research Council of Thailand","keywords":"Medicine; Logistic regression; Stroke (engine); Statistic; Cohort; Proportional hazards model; Retrospective cohort study; Machine learning; Hazard ratio; Naive Bayes classifier; Statistics; Decision tree; Boosting (machine learning); Artificial intelligence; Internal medicine; Computer science; Mathematics; Support vector machine; Confidence interval; Engineering","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":[0.003366129,0.0002699845,0.0008090399,0.0001900015,0.0003173295,0.00004763148,0.0001161814,0.00006117782,0.00003096397],"category_scores_gemma":[0.0004093948,0.0002299172,0.0001523171,0.0003306523,0.00007587719,0.0001212489,0.0001255914,0.0005508433,0.000007398874],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004272492,"about_ca_system_score_gemma":0.00006987584,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008119329,"about_ca_topic_score_gemma":0.000004166065,"domain_scores_codex":[0.997304,0.0003778681,0.0003462096,0.0007048543,0.0008041253,0.0004629404],"domain_scores_gemma":[0.9985982,0.0005903899,0.0001587249,0.000361382,0.0001564896,0.0001348225],"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":[0.0006385826,0.0002200628,0.9430475,0.0001280569,0.00226609,0.00004732072,0.001329953,0.04447974,0.00004591046,0.0007176732,0.002357865,0.004721269],"study_design_scores_gemma":[0.002591812,0.001084944,0.3959902,0.00002857324,0.0008799618,0.00001845868,0.004026089,0.5946528,0.00003247991,0.0002060642,0.0003089143,0.0001797936],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4227221,0.00006449259,0.5522678,0.00004813587,0.00007565055,0.004126684,0.0001267089,0.0003132567,0.02025521],"genre_scores_gemma":[0.8706374,0.00000704676,0.1244897,0.00001237549,0.0001944338,0.002689416,0.0004260966,0.00006628662,0.001477278],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.550173,"threshold_uncertainty_score":0.937575,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1788575098194468,"score_gpt":0.3812605876951467,"score_spread":0.2024030778756999,"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."}}