| 数据挖掘——电信客户流失分析预测 |
| |
| 引用本文： | 王维佳,缪柏其,魏国省.数据挖掘——电信客户流失分析预测[J].数理统计与管理,2006,25(4):419-425. |
| |
| 作者姓名： | 王维佳 缪柏其 魏国省 |
| |
| 作者单位： | 1. 中国科学技术大学统计金融系,合肥,230027
2. 神州数码(中国)有限公司,北京,100085 |
| |
| 摘 要： | 从历史数据中挖掘有效信息从而运用到商业决策中已经成为统计的一个重要应用方向。本文主要是对电信行业客户流失情况进行数据挖掘,针对某运营商的历史数据资料,通过对已流失的客户和在网客户的自然属性和行为属性进行统计分析、挖掘,建立客户流失的预测模型,模型分析导致客户流失的主要因素,预测在网客户在一定时间内流失的可能性。本文中采用IBM的db2平台中的IM(Intelligent Miner)中的决策树和径向基函数进行建模和测试模型,再用IM Scoring进行预测。
|
| 关 键 词： | 客户流失 数据挖掘 决策树 径向基函数 IM IMScoring |
| 文章编号： | 1002-1566(2006)04-0419-07 |
| 收稿时间： | 2005-09-16 |
| 修稿时间： | 2005-11-16 |
Data Mining——Analysis and Predict the Loss of Clients in Telecom |
| |
| WANG Wei-jia,MIAO Bai-qi,WEI Guo-sheng.Data Mining——Analysis and Predict the Loss of Clients in Telecom[J].Application of Statistics and Management,2006,25(4):419-425. |
| |
| Authors: | WANG Wei-jia MIAO Bai-qi WEI Guo-sheng |
| |
| Institution: | 1. Department of Statistics and Finanace of USTC, Hefei 2300027; 2. Digital China, Beijing 100085 |
| |
| Abstract: | It has been a crucial applied direction of Statistics that apply the validate information,mining from the history data,to the decision-making of real business.According to the data in a specifically telecom provider's database,by analyzing and mining the natural attribute and action attribute among the clients,lost or not,we set up a model to predict the probability of loss.From this model,we can also get the main reason which lead to loss.Here,we applied the decision tree and RBF in IM(Intelligent Miner) to set up the model,and IM Scoring to predict.IM and IM Scoring are tow parts of the DB2 platform of IBM. |
| |
| Keywords: | IM IM Scoring |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
