Research on bank credit default prediction based on data mining algorithm

Abstract

It is of great importance to identify the potential risks to the bank's loan customers. Based on data mining technology, it is an effective method to classify loan customers by classification algorithm. In this paper, we use Random Forest method, Logistic Regression method, SVM method and other suitable classification algorithms by python to study and analyze the bank credit data set, and compared these models on five model effect evaluation statistics of Accuracy, Recall, precision, F1-score and ROC area. This paper use the data mining classification algorithm to identify the risk customers from a large number of customers to provide an effective basis for the bank's loan approval.