{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01677398,0.001780827,0.002044728,0.002117426,0.0009056869,0.002261571,0.001988012,0.002065313,0.001358803],"category_scores_gemma":[0.03919296,0.0003399696,0.0007411117,0.00286367,0.001301784,0.004917556,0.002183767,0.002673125,0.0008771853],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001966055,"about_ca_system_score_gemma":0.002010308,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005226027,"about_ca_topic_score_gemma":0.003918648,"domain_scores_codex":[0.9947318,0.002372012,0.0003399245,0.0008133863,0.001310556,0.0004322648],"domain_scores_gemma":[0.9734151,0.01930327,0.001219015,0.003082682,0.002130405,0.0008496047],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.006003864,0.002200264,0.04532848,0.0009464308,0.0007422502,0.000244717,0.0004099146,0.377149,0.005766429,0.008006786,0.03276856,0.5204333],"study_design_scores_gemma":[0.00034523,0.001830019,0.0101381,0.0001607338,0.0002288644,0.0003284745,0.0002649198,0.9653298,0.004716343,0.01210997,0.004507498,0.00004003293],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8341175,0.01680444,0.135726,0.002404605,0.0007170959,0.0002547708,0.001715038,0.002277921,0.005982434],"genre_scores_gemma":[0.930651,0.002389703,0.06028626,0.0005788711,0.0002826746,0.000119978,0.003203243,0.0003702623,0.002117943],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01677398,"threshold_uncertainty_score":0.08871037,"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."}}