{"id":"W1987787956","doi":"10.1145/1281192.1281294","title":"Cleaning disguised missing data","year":2007,"lang":"en","type":"article","venue":"","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Missing data; Outlier; Anomaly detection; Computer science; Data mining; Data quality; Anomaly (physics); Statistics; Artificial intelligence; Mathematics; Machine learning; Engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02367138,0.001583832,0.003116724,0.006076169,0.001772748,0.00259503,0.003968583,0.002268518,0.002725424],"category_scores_gemma":[0.1055645,0.0009983825,0.0017879,0.01117565,0.001675124,0.00375765,0.004274694,0.003811353,0.001822125],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008793615,"about_ca_system_score_gemma":0.002722136,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002090497,"about_ca_topic_score_gemma":0.002226278,"domain_scores_codex":[0.9790788,0.006281776,0.00316842,0.004442421,0.005976832,0.001051654],"domain_scores_gemma":[0.8979059,0.04006054,0.0126013,0.03271887,0.01587025,0.0008431244],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003220239,0.0005742772,0.1289643,0.004849431,0.001167699,0.00497454,0.00626053,0.04403291,0.02383056,0.03041189,0.05486888,0.6968449],"study_design_scores_gemma":[0.0004001564,0.00103551,0.07748143,0.002579334,0.00115901,0.01097391,0.005962683,0.3440078,0.1097917,0.256718,0.1891818,0.0007084981],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0593247,0.001659869,0.9249976,0.001050554,0.000762786,0.0002758074,0.006985664,0.003609094,0.00133381],"genre_scores_gemma":[0.4423099,0.001224712,0.5263528,0.001053706,0.0004525523,0.0007502959,0.02403471,0.001169409,0.002651881],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02367138,"threshold_uncertainty_score":0.1251877,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06706517474363113,"score_gpt":0.3292222302198833,"score_spread":0.2621570554762522,"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."}}