基于改进Boosting的决策树信用卡客户流失分析设想

The Credit Customers Churn Analysis Based on Improved Boosting Decision Tree

当前,信用卡的运营收益成为各商业银行的主要收入来源,但其营销竞争已经趋于白热化。因此,信用卡客户流失分析开始受到各商业银行的重视。而由于决策树具有分类精度高、分类器可解释性强的特点,适合用于客户流失分析。因此,该文提出了利用Boosting技术的决策树集成C5.0进行信用卡客户流失分析的方法。并针对信用卡流失数据分布不对称的特点,对Boosting技术进行了相应的改进尝试。

Currently,the earnings of the credit card has become the main source of income for banks,but it’s marketing competition has become intense.Therefore,the credit card customers churn analysis begun to attract the attention of banks.Because of the high classifica tion accuracy and interpretability strong,the decision tree is suitable for the customers churn analysis.So this paper proposes use the boost ing decision tree technology for credit card customers churn analysis.And according to the data asymmetries in the distribution,try to im prove the boosting technology.