{"id":"W3006371311","doi":"","title":"INDEPENDENT DE-DUPLICATION IN DATA CLEANING","year":2005,"lang":"en","type":"article","venue":"University of Zagreb University Computing Centre (SRCE)","topic":"Data Mining Algorithms and Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Windsor; University of Calgary","funders":"","keywords":"Gene duplication; Biology; Genetics; Gene","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003475109,0.0001141958,0.0001579209,0.000219637,0.0002398913,0.00003549488,0.002517655,0.00007642728,0.00002181786],"category_scores_gemma":[0.00001712359,0.0001631931,0.00004215058,0.0005547562,0.00007316657,0.0007655557,0.001755664,0.0001882167,0.00004301538],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002674304,"about_ca_system_score_gemma":0.0001291105,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007417931,"about_ca_topic_score_gemma":0.0005833369,"domain_scores_codex":[0.9987841,0.000080281,0.0001266691,0.0005217909,0.0001997291,0.0002874657],"domain_scores_gemma":[0.9986182,0.00009765046,0.000170704,0.0009223059,0.00007954501,0.0001115276],"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":[0.00004881836,0.0008888743,0.00478738,0.00005499839,0.0001031966,0.0001770982,0.1932155,0.01406121,0.0007142284,0.06221789,0.008172465,0.7155584],"study_design_scores_gemma":[0.0008385045,0.00001512905,0.01338156,0.00006636029,0.00002060598,0.00001147082,0.03956123,0.9322436,0.00007365588,0.00004548639,0.01350343,0.0002388958],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3935786,0.00001713881,0.6033906,0.001740697,0.00003888047,0.0001153884,0.00003234474,0.000135829,0.0009505629],"genre_scores_gemma":[0.8437744,0.00002542107,0.1555068,0.00003832425,0.00002747353,3.849657e-12,0.00005444479,0.000004104152,0.0005689873],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9181824,"threshold_uncertainty_score":0.665482,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01565594018702825,"score_gpt":0.2107606883378018,"score_spread":0.1951047481507736,"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."}}