{"id":"W1995541812","doi":"10.1007/s00778-009-0161-2","title":"Creating probabilistic databases from duplicated data","year":2009,"lang":"en","type":"article","venue":"The VLDB Journal","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":61,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Probabilistic logic; Data mining; Scalability; Data deduplication; Task (project management); Cluster analysis; Database; String (physics); Probabilistic database; Information retrieval; Machine learning; Artificial intelligence; Database theory; 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":[0.011814,0.001015899,0.002131192,0.005695032,0.001478626,0.006886034,0.006783619,0.002236261,0.003579563],"category_scores_gemma":[0.07379757,0.002510847,0.002979721,0.00944797,0.001422532,0.0117748,0.007296313,0.002279842,0.001551984],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001218541,"about_ca_system_score_gemma":0.003040248,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002422068,"about_ca_topic_score_gemma":0.003470686,"domain_scores_codex":[0.9881392,0.002553105,0.001274037,0.002108176,0.005480946,0.0004445137],"domain_scores_gemma":[0.937184,0.0342419,0.002999665,0.01905036,0.005741648,0.0007822549],"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":[0.001034216,0.0004380929,0.01609243,0.001047934,0.0007365667,0.002295628,0.001905103,0.1630905,0.01249567,0.07947756,0.02216232,0.6992239],"study_design_scores_gemma":[0.0001079395,0.0001601663,0.001609265,0.0001122813,0.0003219376,0.001748828,0.0005635697,0.8275191,0.02251518,0.1282526,0.01696126,0.0001278225],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0336339,0.0003936307,0.9544981,0.0005678157,0.00009633327,0.0002014845,0.001916215,0.007510467,0.001182081],"genre_scores_gemma":[0.1832422,0.000406788,0.8093007,0.000148443,0.00009252705,0.0002299795,0.004675475,0.0009120769,0.0009918391],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.011814,"threshold_uncertainty_score":0.06247914,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4798807441543463,"score_gpt":0.4770797275385156,"score_spread":0.002801016615830709,"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."}}