Credit Scoring Model based on Kernel Density Estimation and Support Vector Machine for Group Feature Selection

A credit scoring model (CSM) is a tool that is typically used in the decision-making process of accepting or rejecting a loan. The selection of an appropriate feature subset is crucial for the credit scoring model. In this paper, we propose a novel framework to improve the performance of this task. First, the kernel density estimation (KDE) is used to construct feature groups in order to combine similar features and reduce wasteful computational workload. Second, the correlation among features is not only simply similar, but also other meaningful relations, such as part-of, has-a etc. Therefore, we calculated the corresponding group scores for each feature group, and then obtained the corresponding radar map according to the group score. The purpose is to help improve the quality of the final selected feature subset and to get the specific semantics of each feature group. Finally, each feature group is selected as a separate entity for feature selection to obtain the optimal feature subset. All features are treated as one-dimensional vectors. The support vector machine (SVM) algorithm is used for training and prediction, and corresponding calculations are performed to obtain a total credit score. Extensive experiments on the UCI benchmark database show the advantages and effectiveness of our proposed algorithm.