Various statistical techniques and machine learning have been used to develop financial prediction models. In this case, credit rating is closely related in terms of prediction of creditworthiness. Because there is no general agreement on financial ratios as an input feature for model development, many studies consider feature selection as a pre-consideration step in data mining before creating a model. This study examines the effect of feature selection using Artificial Bee Colony on the performance improvement of the CART algorithm. The experimental results show that ABC is the best combination of feature selection in improving CART algorithm performance. Compared with some of the proposed PSO and Ant Colony optimization algorithms, this research is expected to be a reference in terms of credit scoring, supporting banks to reject prospective borrowers with poor creditworthiness.
Copyrights © 2021