An improved support vector machine and its application in P2P lending personal credit scoring

Abstract

With the help of Internet technologies, P2P(Peer-to-Peer) lending industry has witnessed the rapid development of loan market. From the reason presented above, credit assessment becomes more and more important to the healthy development of P2P load marked. In order to improve accurate predictions of credit assessment, there is necessary to a kind of credit risk evaluation model based on SVM(Support Vector Machines). The performance of SVM depends, to a great extent, on parameters we chose, therefore, our prior work is optimize them. This paper employs an IFOA(Improved Fruit Fly Optimization algorithm) to optimize parameters of SVM model and uses modified model to analyze P2P load data. In the article, we analyze data with four different ways (Linear Regression, Classical SVM, FOA-SVM and IFOA-SVM), and results show that the one presented in this paper has better accurate predictions.