{"id":"W2950622432","doi":"","title":"Guided Data Repair","year":2011,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Rank (graph theory); Process (computing); Quality (philosophy); Set (abstract data type); Data quality; Data mining; Data set; Machine learning; Information retrieval; Artificial intelligence; 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":[0.02086793,0.001814936,0.002282661,0.002650017,0.001530854,0.003278563,0.007545523,0.002838955,0.005041622],"category_scores_gemma":[0.09035598,0.001257402,0.00186209,0.001813747,0.002009061,0.006051117,0.006241383,0.003899954,0.002691417],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001109289,"about_ca_system_score_gemma":0.002726263,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003180991,"about_ca_topic_score_gemma":0.004462642,"domain_scores_codex":[0.9756905,0.009785255,0.001849265,0.004323745,0.007567562,0.0007836528],"domain_scores_gemma":[0.879153,0.05274453,0.005978447,0.05141124,0.009673699,0.001039074],"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":[0.001336817,0.0008170035,0.01005138,0.001406575,0.0003245325,0.0005594119,0.003225004,0.0550594,0.02533982,0.01490025,0.02662761,0.8603522],"study_design_scores_gemma":[0.0003104632,0.001216293,0.005065637,0.0005178339,0.000252032,0.003416961,0.002209997,0.7130802,0.09545797,0.06634092,0.1117528,0.000378996],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01703925,0.0007738121,0.96743,0.0006393586,0.0001264779,0.000393018,0.0004823343,0.01182736,0.001288401],"genre_scores_gemma":[0.1859119,0.000299231,0.8068439,0.0007627778,0.00008549108,0.000309734,0.001778014,0.001375786,0.002633147],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02086793,"threshold_uncertainty_score":0.1103615,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.7109900226280569,"score_gpt":0.3369640349040057,"score_spread":0.3740259877240512,"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."}}