{"id":"W1991724805","doi":"10.1186/1472-6947-11-53","title":"De-identifying a public use microdata file from the Canadian national discharge abstract database","year":2011,"lang":"en","type":"article","venue":"BMC Medical Informatics and Decision Making","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":31,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Health Information; Agricultural Research Institute of Ontario","funders":"Canadian Institutes of Health Research; CHEO Research Institute; Ontario Institute for Cancer Research; Google","keywords":"Microdata (statistics); Health informatics; Computer science; Data quality; Data mining; Data set; Identification (biology); Data file; Information retrieval; Data collection; Database; Set (abstract data type); Public health; Data science; Medicine; Statistics; Population; Census; Environmental health; Artificial intelligence; Engineering","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00579258,0.0006879446,0.0005102684,0.005383748,0.001715149,0.001533555,0.001976381,0.0006338391,0.001749652],"category_scores_gemma":[0.04167763,0.0003288522,0.00041507,0.008155041,0.0007457873,0.0008382502,0.001566342,0.0008533144,0.0005549949],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007242648,"about_ca_system_score_gemma":0.0158894,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.4558502,"about_ca_topic_score_gemma":0.42986,"domain_scores_codex":[0.9887504,0.0006624328,0.0009357099,0.0009706405,0.007917338,0.0007633991],"domain_scores_gemma":[0.9683026,0.008703306,0.005095746,0.005313595,0.01206918,0.0005154746],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002101979,0.000427357,0.3907733,0.001534073,0.0004102267,0.001530106,0.00387995,0.03967906,0.01550776,0.01320306,0.1551304,0.3758228],"study_design_scores_gemma":[0.000181827,0.0002715229,0.497293,0.0003633979,0.0002239694,0.001297619,0.002697547,0.2789237,0.05598809,0.004484799,0.1580318,0.0002426862],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6927353,0.0004781126,0.0546965,0.003008012,0.0002401909,0.00327004,0.2217925,0.008466081,0.01531329],"genre_scores_gemma":[0.7490348,0.0003040919,0.08723026,0.0002896047,0.00005915482,0.001126158,0.1575997,0.0002199793,0.004136286],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5441498,"threshold_uncertainty_score":0.9063937,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5616257091682858,"score_gpt":0.4819889484445266,"score_spread":0.07963676072375925,"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."}}