Extreme Learning Machine Enhanced Gradient Boosting for Credit Scoring

Credit scoring is an effective tool for banks and lending companies to manage the potential credit risk of borrowers. Machine learning algorithms have made grand progress in automatic accurate discrimination good bad Notably, ensemble approaches are a group powerful tools enhance performance scoring. Random forest (RF) Gradient Boosting Decision Tree (GBDT) become mainstream methods precise RF Bagging-based that realizes enriches diversity base learners by modifying training object. However, optimization pattern works on invariant targets may increase statistical independence learners. GBDT boosting-based approach reduces error iteratively changing target while keeping features unchanged. This harm In this study, we incorporate advantages Bagging strategy boosting An extreme machine-based supervised augmented proposed discriminative ability Experimental results 4 public datasets show significant improvement suggest method solution realize