Credit scoring model based on SyMProD-ENN and logistic regression

Unbalanced data distribution is a common problem in the development process of credit scoring cards. This paper proposes a new method of mixed sampling, SyMProD-ENN, to improve the problem of developing effective credit scoring cards on imbalanced datasets. The experimental results show that the SyMProD-ENN method can effectively improve the classification performance of the credit scoring model constructed based on the logistic regression methods, especially suitable for credit data with imbalance category, high-dimensional characteristics, non-normal distribution and outliers.