{"id":"W3216023068","doi":"10.1108/dta-05-2021-0108","title":"Data repair of density-based data cleaning approach using conditional functional dependencies","year":2021,"lang":"en","type":"article","venue":"Data Technologies and Applications","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University","funders":"","keywords":"Computer science; Data mining; Data quality; Functional dependency; Scalability; Set (abstract data type); Data set; Quality (philosophy); Data validation; Database; Artificial intelligence; Engineering; Relational database","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":[0.00478875,0.001031816,0.001229875,0.003769745,0.001577219,0.001382073,0.002540278,0.0009209772,0.001387923],"category_scores_gemma":[0.01976831,0.0006658004,0.00212071,0.002538906,0.001149117,0.002393816,0.002791052,0.001567402,0.0005881265],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001367598,"about_ca_system_score_gemma":0.003441386,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0113169,"about_ca_topic_score_gemma":0.01056317,"domain_scores_codex":[0.9932724,0.001154039,0.0005836463,0.001408491,0.003244611,0.0003367771],"domain_scores_gemma":[0.9809497,0.007068116,0.001715965,0.003905069,0.006085452,0.0002757179],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005319115,0.0004751836,0.03510231,0.0009306985,0.0003888783,0.0009429512,0.001344499,0.2236009,0.03977846,0.01323279,0.009129209,0.6745422],"study_design_scores_gemma":[0.00003708039,0.0002389064,0.01124696,0.0001040612,0.0002113034,0.001058036,0.0003641766,0.8978255,0.06384207,0.01396333,0.01100856,0.0001000721],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03284377,0.0003273041,0.9624413,0.0002483237,0.00004565985,0.0001719768,0.0003606472,0.002862523,0.0006984379],"genre_scores_gemma":[0.3591113,0.0002984806,0.6362333,0.0002305241,0.00004507612,0.0002770407,0.002086863,0.0003647299,0.001352653],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0113169,"threshold_uncertainty_score":0.02532566,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5594296950545075,"score_gpt":0.4357612867051186,"score_spread":0.1236684083493889,"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."}}