摘要
针对电信领域客户流失的问题,提出了改进聚类的客户流失预测模型。根据通信行业中实际客户流失数据的正负样本数量不平衡而且数据量特别大的特点,提出带有不同权重参数的改进聚类算法,并将其用于电信行业的客户流失预测模型中。通过实际电信客户数据集测试,与传统的预测算法比较,证明这种算法适合解决大数据集和不平衡数据,具有更高的精确度,能够取得较好的客户流失预测效果。
The analysis of telecom client churn is the main content of the client relationship management in communication industry.Aiming at the problem of client churn,a client churn prediction model based on the improved clustering algorithm is presented.According to the specific characteristics of the large data amount and imbalance between the quantity of positive and negative samples in actual client churn,an improved clustering algorithm with different weight parameters is proposed and applied to the client churn prediction model in telecommunication industry.Comparison between the results tested by proposed algorithm and the traditional prediction algorithm for the data set of actual telecom client shows that the proposed algorithm is suitable for solving large-scale data set and imbalance data.Besides,the proposed algorithm has an advantage in high precision and can achieve a good prediction result for the client churn.
出处
《太原理工大学学报》
CAS
北大核心
2014年第4期532-536,共5页
Journal of Taiyuan University of Technology
基金
国家自然科学基金资助项目(61301250)
关键词
聚类
客户流失
加权
预测分析
clustering
client churn
weighted
prediction analysis