{"id":"W2165739648","doi":"10.1109/icpc.2008.41","title":"NICAD: Accurate Detection of Near-Miss Intentional Clones Using Flexible Pretty-Printing and Code Normalization","year":2008,"lang":"en","type":"article","venue":"","topic":"Software Engineering Research","field":"Computer Science","cited_by":519,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Computer science; Parsing; Normalization (sociology); Agile software development; Programming language; Precision and recall; Transformation (genetics); Source code; Disk formatting; Code generation; Rule-based machine translation; Compiler; Code (set theory); Artificial intelligence; Natural language processing; Software engineering; Operating system","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":[0.002086135,0.0008835479,0.001216416,0.002579856,0.0007649033,0.001812474,0.002108786,0.001148836,0.001463324],"category_scores_gemma":[0.01376857,0.0005228861,0.000663281,0.00170494,0.0009638674,0.002254108,0.001694474,0.001154388,0.001125936],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001064942,"about_ca_system_score_gemma":0.001732855,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002271197,"about_ca_topic_score_gemma":0.003468752,"domain_scores_codex":[0.9966757,0.0005355753,0.0002444454,0.0007269615,0.001684431,0.0001328905],"domain_scores_gemma":[0.9915549,0.003250394,0.0009122303,0.002206533,0.001920774,0.000155063],"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":[0.0005310707,0.0001565734,0.01954482,0.0003560561,0.00009840062,0.0005238942,0.0007575767,0.01323293,0.119208,0.00791916,0.006682642,0.8309889],"study_design_scores_gemma":[0.00008508107,0.0002986426,0.013542,0.00007692396,0.000171523,0.003339975,0.0002826367,0.4941291,0.4468942,0.009192022,0.03175686,0.000230951],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05886842,0.0004647537,0.9097724,0.0001527233,0.00007281176,0.00011728,0.0002610371,0.02815305,0.002137561],"genre_scores_gemma":[0.1743131,0.0001850573,0.8189465,0.0001583313,0.00002461302,0.0001580477,0.0009552276,0.001513551,0.003745397],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002579856,"threshold_uncertainty_score":0.01103264,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0470208530369328,"score_gpt":0.2932180374607296,"score_spread":0.2461971844237968,"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."}}