{"id":"W3209614841","doi":"10.1007/s00799-021-00311-0","title":"Improving data quality in large-scale repositories through conflict resolution","year":2021,"lang":"en","type":"article","venue":"International Journal on Digital Libraries","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Technische Universität Wien; Vienna Science and Technology Fund; Technische Universität Wien Bibliothek; Ontario Research Foundation","keywords":"Metadata; Computer science; Data quality; Correctness; Scalability; Profiling (computer programming); Data mining; Data science; Conflict resolution; Quality (philosophy); Data collection; Database; Information retrieval; World Wide Web","routes":{"ca_aff":true,"ca_fund":true,"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":[0.03979885,0.0009320941,0.001273028,0.006338681,0.001847234,0.006116795,0.003651687,0.00136996,0.0009155959],"category_scores_gemma":[0.1123047,0.0006170771,0.001128711,0.004724098,0.00179536,0.004853058,0.004918857,0.001648317,0.0003923188],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001720713,"about_ca_system_score_gemma":0.00328097,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002759299,"about_ca_topic_score_gemma":0.002705838,"domain_scores_codex":[0.94802,0.02406392,0.005376717,0.004132052,0.01738762,0.001019802],"domain_scores_gemma":[0.8085667,0.1027922,0.02345755,0.03435808,0.02830411,0.002521312],"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":[0.0006941844,0.001751097,0.09229024,0.0008942829,0.0005151966,0.0007510141,0.003358458,0.1738154,0.03832098,0.01789358,0.005231797,0.6644838],"study_design_scores_gemma":[0.0001157917,0.0004295461,0.01165833,0.0002365177,0.0001850958,0.0005642017,0.001379717,0.8561876,0.09252352,0.0283066,0.008262184,0.0001508226],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1494925,0.0004154952,0.8431202,0.0007962265,0.00004148295,0.0006618107,0.0003491059,0.002789729,0.002333391],"genre_scores_gemma":[0.4036821,0.0001087779,0.594833,0.0001499372,0.00002294396,0.000169819,0.0004934384,0.0001351106,0.0004048482],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03979885,"threshold_uncertainty_score":0.2104789,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2067003591168642,"score_gpt":0.4259471446766152,"score_spread":0.219246785559751,"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."}}