{"id":"W2602793737","doi":"10.1109/msr.2017.50","title":"Rediscovery Datasets: Connecting Duplicate Reports","year":2017,"lang":"en","type":"article","venue":"","topic":"Software Engineering Research","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Software bug; Open source software; Software; Open source; Tracking (education); Tracking 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005034702,0.00007157079,0.00008299045,0.00004450923,0.0003341748,0.0007959878,0.00129915,0.00002907133,0.00001762776],"category_scores_gemma":[0.00278084,0.00006335074,0.00002571495,0.00005406943,0.00002775593,0.001000717,0.001121223,0.0001154459,0.00007464664],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002337509,"about_ca_system_score_gemma":0.00003483254,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001453761,"about_ca_topic_score_gemma":0.000005737299,"domain_scores_codex":[0.9990091,0.000008456652,0.0001345909,0.0003670541,0.0002354212,0.0002453975],"domain_scores_gemma":[0.9968016,0.0002347537,0.00008218402,0.002764713,0.00002939514,0.00008741803],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.0000150377,0.0002422223,0.2080232,0.0001785607,0.0001894379,0.008702103,0.0006817235,0.001600935,0.01398202,0.07477749,0.3504943,0.341113],"study_design_scores_gemma":[0.001186065,0.0002352252,0.3949189,0.0002192615,0.00001591251,0.003238143,0.00005559605,0.105071,0.09305224,0.01557608,0.3843436,0.002087951],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0600712,0.00004120383,0.9276561,0.001971506,0.001004102,0.0001533166,0.00000772023,0.0008142623,0.008280658],"genre_scores_gemma":[0.9620646,0.000004332132,0.03743136,0.00005659241,0.00007962793,0.000007721785,0.00000878132,0.000008463712,0.0003385117],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9019934,"threshold_uncertainty_score":0.7675729,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03030703116912421,"score_gpt":0.3088519965327439,"score_spread":0.2785449653636197,"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."}}