{"id":"W4321480056","doi":"10.1055/s-0043-1762904","title":"Evaluation of Feature Selection Methods for Preserving Machine Learning Performance in the Presence of Temporal Dataset Shift in Clinical Medicine","year":2023,"lang":"en","type":"article","venue":"Methods of Information in Medicine","topic":"Sepsis Diagnosis and Treatment","field":"Medicine","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute for Clinical Evaluative Sciences; SickKids Foundation; Hospital for Sick Children","funders":"","keywords":"Feature selection; Logistic regression; Artificial intelligence; Computer science; Oracle; Feature (linguistics); Machine learning; Predictive modelling; Model selection; Feature engineering; Statistics; Deep learning; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.06118327,0.0001538841,0.0008014672,0.0009473077,0.0000302144,0.000002735054,0.0002109737,0.0001443177,0.00006902622],"category_scores_gemma":[0.02260583,0.0000946398,0.00005950796,0.001728961,0.0001683752,0.0003706827,0.00005205669,0.0004158129,5.977649e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008295984,"about_ca_system_score_gemma":0.0001325591,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007036842,"about_ca_topic_score_gemma":0.0001457768,"domain_scores_codex":[0.9941424,0.002570049,0.001976599,0.0001821307,0.0009158702,0.000212976],"domain_scores_gemma":[0.9937334,0.004655381,0.0008146759,0.0003275575,0.0004277715,0.00004122537],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0005893318,0.0002273939,0.6422295,0.001341632,0.00007725322,6.765609e-7,0.009840923,0.005502935,0.0003545574,0.0001269526,0.002343028,0.3373658],"study_design_scores_gemma":[0.006371405,0.001404048,0.6578066,0.001369564,0.0002082555,0.000004406127,0.002053149,0.3258946,0.001364812,0.0004701849,0.002989385,0.00006360466],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9179201,0.003044134,0.05499155,0.01546623,0.000672992,0.006408992,0.0001016518,0.00004836552,0.001345987],"genre_scores_gemma":[0.8652479,0.0009764293,0.1311433,0.0002402417,0.000092656,0.0003505803,0.001926817,0.00001215311,0.000009908895],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3373022,"threshold_uncertainty_score":0.9856272,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2786180993514724,"score_gpt":0.570282405381938,"score_spread":0.2916643060304656,"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."}}