{"id":"W4417435774","doi":"10.3390/jrfm18120724","title":"Balancing Fairness and Accuracy in Machine Learning-Based Probability of Default Modeling via Threshold Optimization","year":2025,"lang":"en","type":"article","venue":"Journal of risk and financial management","topic":"Financial Distress and Bankruptcy Prediction","field":"Business, Management and Accounting","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Transparency (behavior); Retraining; Fairness measure; Selection (genetic algorithm); Model selection; Logistic regression; Credit risk","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006744438,0.0001282708,0.0002774741,0.0004185068,0.0001258844,0.00007334902,0.00009853838,0.00006338119,0.000004993009],"category_scores_gemma":[0.0002572211,0.0001136152,0.00005879943,0.0004257363,0.00003570878,0.000480954,0.0001055834,0.0002175057,2.17717e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000322694,"about_ca_system_score_gemma":0.00001987688,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004723905,"about_ca_topic_score_gemma":0.0002072223,"domain_scores_codex":[0.9989685,0.00001640298,0.0005357154,0.0001640217,0.0001751415,0.0001401877],"domain_scores_gemma":[0.9992977,0.00003263068,0.0004163622,0.00008338379,0.0001606645,0.000009305601],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003475875,0.0001521358,0.2214416,0.0007480201,0.00001075516,0.00001111833,0.00004949059,0.6877831,0.00001473974,0.00528841,0.00002237648,0.08413068],"study_design_scores_gemma":[0.001377768,0.00003042022,0.1264213,0.0004182367,0.0001102773,5.955623e-7,0.00007047896,0.861819,0.00001041707,0.008901365,0.0007295376,0.0001106538],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4969184,0.0005127153,0.5018187,0.00008131035,0.0001728841,0.000172551,0.000001943507,0.00001036196,0.0003111975],"genre_scores_gemma":[0.9982182,0.0004597829,0.001117459,0.0000797222,0.0001004766,0.000005253819,0.000005918399,0.000006785847,0.000006400627],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5012998,"threshold_uncertainty_score":0.4633095,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.006605454578361702,"score_gpt":0.2028897047597801,"score_spread":0.1962842501814184,"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."}}