Data-Driven Credit Scoring: Integrating Genetic Algorithms With Ensemble Learning for Multi Class Classification

In the unexpectedly dynamic financial world, credit scoring plays a pivotal role in assessing the creditworthiness of people and agencies. Traditional statistical models, though effective, regularly face obstacles in managing complex and large datasets. This study introduces a unique method to credit score scoring by using integrating genetic algorithms with ensemble learning strategies to improve class accuracy. The research utilizes a dataset of consumer financial and demographic variables to broaden predictive models, mainly the use of XGBoost and Random Forest classifiers. Through hyperparameter optimization using genetic algorithms, the models acquire achieve improvements in overall performance. Further, ensemble techniques, consisting of weighted ensembling and stacking, are hired to leverage the strengths of a couple of models. This research promotes the capability of combining genetic algorithms and superior ensemble techniques in growing strong and correct credit score scoring models, contributing to extra efficient risk control inside the financial zone. The findings suggest large implications for banks and monetary establishments in improving decision-making tactics and customer segmentation.