{"id":"W2149864547","doi":"10.1109/isese.2005.1541846","title":"Cloning by accident: an empirical study of source code cloning across software systems","year":2005,"lang":"en","type":"article","venue":"","topic":"Software Engineering Research","field":"Computer Science","cited_by":79,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Cloning (programming); Computer science; Source code; Domain (mathematical analysis); Set (abstract data type); Open source; Code (set theory); Programming language; Software; Notice; Operating system; World Wide Web; Software engineering","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":[0.04469388,0.0005048992,0.0007570008,0.006315522,0.004293696,0.004078347,0.002977605,0.00298751,0.001539035],"category_scores_gemma":[0.3356162,0.0009236122,0.0005881949,0.006661715,0.008718396,0.01047858,0.005644443,0.003634783,0.0003741282],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002900847,"about_ca_system_score_gemma":0.002446877,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006296062,"about_ca_topic_score_gemma":0.004856825,"domain_scores_codex":[0.922362,0.04818507,0.006684821,0.005788494,0.01505401,0.001925653],"domain_scores_gemma":[0.4033622,0.451457,0.08448038,0.02456995,0.03193745,0.004193028],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003646921,0.0007180385,0.7404389,0.0005670288,0.000166203,0.001776186,0.217898,0.0006790907,0.0009659247,0.003680837,0.0008609385,0.03188409],"study_design_scores_gemma":[0.00007536897,0.002207338,0.6304628,0.0007139504,0.0002203979,0.00732924,0.3220585,0.01143125,0.003376429,0.006444124,0.01544925,0.0002313928],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9951031,0.000273226,0.002910414,0.0002882827,0.000008403322,0.0001040488,0.00005466043,0.00002195126,0.001235839],"genre_scores_gemma":[0.9968803,0.0002000597,0.002290532,0.0001429776,0.00001147406,0.00009408183,0.0001060437,0.00002735432,0.0002471833],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04469388,"threshold_uncertainty_score":0.2363666,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03103695912387,"score_gpt":0.3509950186155593,"score_spread":0.3199580594916893,"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."}}