{"id":"W4249958738","doi":"10.1007/978-0-387-39940-9_2347","title":"Data Deduplication","year":2009,"lang":"en","type":"book-chapter","venue":"Encyclopedia of Database Systems","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Data deduplication; Computer science; 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":[0.002314176,0.001111497,0.0009760518,0.003243823,0.00135778,0.003792933,0.002830541,0.0007060173,0.04056437],"category_scores_gemma":[0.006559975,0.0006726925,0.001058363,0.004770921,0.0006724941,0.004367971,0.0036542,0.00138468,0.0233285],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005243661,"about_ca_system_score_gemma":0.001302047,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004541366,"about_ca_topic_score_gemma":0.0005976619,"domain_scores_codex":[0.9984643,0.0002210867,0.0002127239,0.0002995274,0.0006790255,0.0001233728],"domain_scores_gemma":[0.991915,0.0009228185,0.000214511,0.005155641,0.001657974,0.0001341188],"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":[0.0002156941,0.0001061748,0.000960315,0.001008122,0.00007438083,0.0003641177,0.0003108999,0.002003219,0.008093862,0.03692932,0.2295858,0.720348],"study_design_scores_gemma":[0.00005482162,0.00009753812,0.0009261972,0.0003855637,0.0001091271,0.001952753,0.0001985971,0.01138403,0.05616875,0.03961564,0.889027,0.00008002367],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01841493,0.01431069,0.7351065,0.00352492,0.006454461,0.001906646,0.01622557,0.03455527,0.169501],"genre_scores_gemma":[0.1362274,0.01339005,0.4477179,0.003705636,0.001931111,0.001096772,0.0407468,0.009098731,0.3460857],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.04056437,"threshold_uncertainty_score":0.1357014,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2149054609765482,"score_gpt":0.3946475339549821,"score_spread":0.1797420729784339,"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."}}