Abstract: In customer churn analysis of personal financial services in commercial bank, there are large number of customer samples, and the number of churn samples and that of nonchurn samples are imbalanced. Thus, it is a challenging problem to do the churn analysis. To solve this problem, an integrated method that combines boosting algorithm with costsensitive decision tree is presented. To show the effectiveness of the proposed method, it is applied to a case study. Comparison shows that the proposed method outperforms the other existing ones, such as support vector machine, artificial neural network, and logistic regression.  

Key words: customer churn, data mining, decision tree, boosting algorithm, cost-sensitive learning, personal financial services in commercial bank