The Evaluation Model And Empirical Research Of Customers' Loan Approval Based On Decision Tree And Support Vector Machine Algorithm

With the rapid development of Internet, it affects the various industries, of course, the financial industry is no exception. The Internet constantly in technology, promote the development of the financial sector, channels and services step by step changing the future of the financial sector. Currently produced-financial institutions, innovation of Internet financial six forms the Internet platform, fund sales based on Internet, Internet payment, lending to the P2P network, P2P network of microfinance and suggests. In this paper, the financial institution innovation in vertical search platform of the Internet platform in financial research.The lending financial needs of the users of vertical search, user basic information and financial institutions to provide financial products matching and orders for examination and approval. The problems with the asymmetric information problems in the process of loan, matching the information efficiency and risk control problems. This is also this article dedicated to solving the problem, this article mainly aims at improve the efficiency of the examination and approval of order and the correctness of the examination and approval.First work on the decision tree, logistic regression and support vector machine (SVM) algorithm of theoretical study, and rise to the application layer to the understanding and comparing. This article from the Internet for the melting of 360 credit approval data combined with the theoretical background for empirical research of the algorithm. After data collection, data preprocessing and training sample selection, respectively based on the decision tree, logistic regression and support vector machine (SVM) model to predict loan approval, which according to the actual demand, adjust the model parameters.Comparing three kinds of prediction algorithm for classification of empirical research, the loan approval prediction model based on decision tree method whole model is relatively ideal, it has higher precision, lower the risk of cost, strong controllability and high efficiency. This further illustrates the loan approval prediction model based on decision tree method in actual use vertical platform for financial credit matching has certain guiding effect to a certain extent, solves the user and the information asymmetry problem between financial institutions, improve the efficiency of order approval, to reduce the risk and cost of the examination and approval of the wrong, can provide effective and efficient support for the credit decisions.