{"id":"W2565966619","doi":"10.1109/cbi.2016.9","title":"ELM: An Extended Logic Matching Method on Record Linkage Analysis of Disparate Databases for Profiling Data Mining","year":2016,"lang":"en","type":"article","venue":"","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Computer science; Data mining; Probabilistic logic; Profiling (computer programming); Data deduplication; Matching (statistics); Identifier; Artificial intelligence; Machine learning; Database; Mathematics; Statistics","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":[0.01149936,0.0009390591,0.001311995,0.01057228,0.001442255,0.003362615,0.003166933,0.001625989,0.004995164],"category_scores_gemma":[0.03398549,0.0005237712,0.001871143,0.01060861,0.001046872,0.004431249,0.004257646,0.001687939,0.001616674],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00106944,"about_ca_system_score_gemma":0.002319912,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0018212,"about_ca_topic_score_gemma":0.002107152,"domain_scores_codex":[0.9889776,0.003896498,0.001274014,0.001614903,0.003947326,0.0002896693],"domain_scores_gemma":[0.990042,0.00582088,0.001155306,0.001583187,0.001179826,0.0002186949],"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":[0.0005150264,0.0003808053,0.009708775,0.000558315,0.000483362,0.0007384578,0.0007316735,0.02566437,0.007295142,0.0472733,0.007867398,0.8987834],"study_design_scores_gemma":[0.0001260682,0.0003122169,0.00469426,0.0002458073,0.0002501616,0.00161891,0.0004609632,0.7711545,0.013468,0.1780024,0.02952492,0.0001418494],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.005600905,0.0002113149,0.9894387,0.0001774389,0.0000353972,0.0002549318,0.0006019323,0.0027818,0.0008975209],"genre_scores_gemma":[0.05492419,0.0001401005,0.942147,0.0001653475,0.00004512178,0.0003032523,0.001301485,0.0001798739,0.0007936445],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01149936,"threshold_uncertainty_score":0.06081516,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5979240502219576,"score_gpt":0.5597760552806768,"score_spread":0.03814799494128085,"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."}}