{"id":"W4408328303","doi":"10.1016/j.neucom.2025.129896","title":"Improving GBDT performance on imbalanced datasets: An empirical study of class-balanced loss functions","year":2025,"lang":"en","type":"article","venue":"Neurocomputing","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":29,"is_retracted":false,"has_abstract":false,"ca_institutions":"Fields Institute for Research in Mathematical Sciences","funders":"","keywords":"Computer science; Class (philosophy); Artificial intelligence; 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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0004322447,0.0002591075,0.0003379077,0.0003472969,0.0003561064,0.0001567913,0.001723655,0.00008334823,0.000002513884],"category_scores_gemma":[0.0001357799,0.0002558646,0.00004905894,0.001200353,0.00006496031,0.0007354541,0.000660631,0.0004412338,0.00001407875],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008399267,"about_ca_system_score_gemma":0.0001146868,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002220418,"about_ca_topic_score_gemma":0.000005991226,"domain_scores_codex":[0.9973773,0.0002103497,0.0006425804,0.0009571532,0.0004037475,0.0004088811],"domain_scores_gemma":[0.9973031,0.0002356687,0.0003476246,0.001891682,0.0001370207,0.00008496273],"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":[0.000240752,0.006893073,0.2392457,0.0003528464,0.0001026115,0.00005136612,0.001646852,0.007426241,0.05711972,0.005898232,0.009773106,0.6712494],"study_design_scores_gemma":[0.001883349,0.002415825,0.2836084,0.0001446377,0.00002497714,0.00001469322,0.0002274554,0.6889524,0.01718351,0.0001277719,0.004906269,0.0005106755],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7101345,0.000006459167,0.2873063,0.0002288639,0.000422087,0.0005579776,0.00003064947,0.0007193486,0.0005938407],"genre_scores_gemma":[0.9896443,0.00000315764,0.009377777,0.000733894,0.00006386371,0.00004862541,0.00006488258,0.0000160156,0.00004751317],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6815262,"threshold_uncertainty_score":0.9999893,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02398388957571029,"score_gpt":0.3180405051427695,"score_spread":0.2940566155670592,"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."}}