Decision Tree Application to Classification Problems with Boosting Algorithm

A personal credit evaluation algorithm is proposed by the design of a decision tree with boosting algorithm, and classification carried out. By comparison conventional it shown that acts to speed up processing time. The Classification Regression Tree (CART) showed 90.95% accuracy, slightly higher than without boosting, 90.31%. To avoid overfitting model on training set due unreasonable data division, we consider cross-validation illustrate results simulation; hypermeters have been applied fitting effect verified. fitted optimally help confusion matrix. In this paper, relevant indicators are also introduced evaluate performance model. For methods, accuracy rate, error precision, recall, etc. illustrated; comprehensively based after 10-fold cross-validation. show improves in precision when CART applied, but time takes much longer, around 2 min. With obtained result, verified improved under algorithm. At same time, test verification fitting, could be prediction for customers’ decisions subscription fixed deposit business.