A Practical Feature Selection Approach for Customer Churn Prediction in Telecommunications

摘要：

There is little literature to introduce the approaches for the feature selection, which plays an important role in the customer churn prediction. In addition, due to the imbalanced data sets where the number of majority class (nonchurn) is much higher than the minority class (churner), most of the traditional feature selection approaches is ineffective for the churn prediction. This paper proposes a new practical filter feature selection approach for customer churn prediction in telecom- munications service field. Based on Chi-square and the standard residual analysis methods, this approach calculates the dependency between each input feature and the classes for selecting features. Finally, the comparative experiments were carried out and the experimental results show that the new proposed feature selection approach is effective for the churn prediction.

关键词： churn prediction feature selection significant level

作者: B. Q. Huang Y. Huang M. T. Kechadi

作者单位: School of Computer Science and Informatics, University College Dublin, Dublin, Ireland

会议类型: 国际会议

会议名称: 2010 International Conference on Future Information Technology(2010年未来信息技术国际会议 ICFIT 2010)

会议地点: 长沙

会议语种:英文

页码: 126-130

在线出版日期: 2010-12-14（万方平台首次上网日期，不代表论文的发表时间）