{"id":"W2107080109","doi":"10.14778/1454159.1454220","title":"A revival of integrity constraints for data cleaning","year":2008,"lang":"en","type":"article","venue":"Proceedings of the VLDB Endowment","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":46,"is_retracted":false,"has_abstract":true,"ca_institutions":"Bell (Canada)","funders":"Engineering and Physical Sciences Research Council","keywords":"Data integrity; Data quality; Computer science; Schema (genetic algorithms); Constraint (computer-aided design); Quality (philosophy); Risk analysis (engineering); Data science; Reliability engineering; Database; Information retrieval; Engineering; Operations management; Business; Mechanical 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":[0.06524403,0.00179881,0.003985106,0.005032946,0.003915077,0.01395734,0.006697068,0.00608074,0.004544493],"category_scores_gemma":[0.126332,0.00213287,0.003984455,0.01032983,0.01051364,0.03631319,0.01204255,0.02394943,0.003357954],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005809441,"about_ca_system_score_gemma":0.006911201,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004303015,"about_ca_topic_score_gemma":0.00280382,"domain_scores_codex":[0.931708,0.02556066,0.005733217,0.006661665,0.02890004,0.001436318],"domain_scores_gemma":[0.8320317,0.0923043,0.004234839,0.03807764,0.03088335,0.002468228],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001827246,0.0000877034,0.0008813656,0.002355304,0.0002699356,0.0002744001,0.001425924,0.004596581,0.003450118,0.5161628,0.07784501,0.3924681],"study_design_scores_gemma":[0.00005291304,0.00009399231,0.0006463315,0.00127854,0.0001034419,0.0008784162,0.0006140263,0.02369583,0.007072672,0.4366288,0.5287421,0.0001927813],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00282829,0.05926521,0.859722,0.06505173,0.003205241,0.0001516559,0.0007679772,0.001377381,0.00763064],"genre_scores_gemma":[0.04393539,0.06594352,0.8588219,0.01631306,0.006247823,0.0003991133,0.00169559,0.001391892,0.005251735],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.06524403,"threshold_uncertainty_score":0.3450475,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5397355688626213,"score_gpt":0.4498791965858916,"score_spread":0.08985637227672966,"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."}}