{"id":"W4400685466","doi":"10.1145/3676961","title":"SimClone: Detecting Tabular Data Clones Using Value Similarity","year":2024,"lang":"en","type":"article","venue":"ACM Transactions on Software Engineering and Methodology","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University; Huawei Technologies (Canada); University of Manitoba","funders":"","keywords":"Computer science; Data mining; Header; Software; Visualization; clone (Java method); Similarity (geometry); Margin (machine learning); Artificial intelligence; Image (mathematics); Machine learning","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":[0.00435314,0.001755558,0.00157233,0.008597459,0.001106738,0.002961926,0.002304763,0.001876571,0.00181507],"category_scores_gemma":[0.03068519,0.0005044657,0.001615589,0.006905839,0.0007223578,0.004066015,0.002476622,0.001293641,0.002242107],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008359987,"about_ca_system_score_gemma":0.00158952,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002969858,"about_ca_topic_score_gemma":0.004724973,"domain_scores_codex":[0.9939301,0.0009319231,0.0008229218,0.001686845,0.002382433,0.0002458499],"domain_scores_gemma":[0.9758462,0.009042843,0.004354957,0.005008091,0.005215513,0.0005323804],"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":[0.001213809,0.0007249763,0.2219347,0.001390245,0.000863576,0.00115666,0.001816645,0.01742617,0.04858864,0.005800591,0.06149435,0.6375897],"study_design_scores_gemma":[0.0001766416,0.0007853775,0.05326982,0.0002360288,0.0002528843,0.002438635,0.001089116,0.7721716,0.114144,0.01549726,0.03968674,0.0002519049],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3299357,0.003010132,0.5207184,0.0008367124,0.0004027587,0.0009040813,0.02239016,0.1171844,0.004617651],"genre_scores_gemma":[0.5165632,0.0004744499,0.4413514,0.0004726313,0.0001256408,0.0006241468,0.03522575,0.002573773,0.002588956],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008597459,"threshold_uncertainty_score":0.02302188,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.549696504884526,"score_gpt":0.4849508565192975,"score_spread":0.06474564836522856,"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."}}