{"id":"W4224288101","doi":"10.1145/3524303","title":"Contextual Data Cleaning with Ontology Functional Dependencies","year":2022,"lang":"en","type":"article","venue":"Journal of Data and Information Quality","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Tech University; University of Waterloo; McMaster University","funders":"","keywords":"Ontology; Axiom; Computer science; Functional dependency; Inference; Dependency (UML); Set (abstract data type); Dependency theory (database theory); Relation (database); Data mining; Artificial intelligence; Relational database; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01280239,0.001243768,0.001304515,0.0050879,0.002656621,0.003263409,0.003442183,0.001836035,0.001586263],"category_scores_gemma":[0.0736202,0.001182292,0.003740099,0.005494535,0.003134922,0.006782784,0.007689937,0.00373953,0.0004864841],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002123281,"about_ca_system_score_gemma":0.006386623,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01121266,"about_ca_topic_score_gemma":0.01602696,"domain_scores_codex":[0.9794614,0.00682774,0.001962819,0.004136915,0.006737336,0.0008737772],"domain_scores_gemma":[0.9316756,0.03637591,0.005555235,0.01868323,0.007211703,0.0004982552],"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":[0.0004681487,0.0004289915,0.04186444,0.002014238,0.0006958778,0.001640723,0.003385196,0.1520008,0.02277221,0.1707439,0.01503735,0.588948],"study_design_scores_gemma":[0.00007202521,0.0002059869,0.007526405,0.0004824576,0.0004333254,0.001647893,0.001416314,0.5662795,0.05752672,0.3157865,0.04842591,0.0001969989],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01189464,0.0002454798,0.9846015,0.0005385164,0.00003807854,0.0001720083,0.0005525779,0.001190898,0.0007663086],"genre_scores_gemma":[0.1265616,0.0002376093,0.8702582,0.0003603526,0.00005106887,0.0001964058,0.001461933,0.0002612167,0.0006116922],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01280239,"threshold_uncertainty_score":0.06770635,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4677702605003999,"score_gpt":0.452390842963974,"score_spread":0.01537941753642585,"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."}}