数据挖掘技术在电信客户流失预警系统中的研究  

Studies of the data mining technology in the telecom customer churn warning system

作　　者：薛素静[1] 陈帆[1] 

机构地区：[1]华北水利水电学院,河南郑州450011

出　　处：《微型机与应用》2011年第21期49-52,共4页Microcomputer & Its Applications

摘　　要：在明确业务问题的基础上,筛选有效的输入数据和目标变量,并对输入变量各参数之间以及输入变量与目标变量之间进行相关性分析,选取有效的参数。在数据准备完成的基础上,利用Neural Network来建立预测模型,并给出预测结果,通过运行实际业务中的数据对模型进行评估。通过该模型预测可能流失的客户,并给出预警信号,以便企业做出经营决策,挽留有关用户,确保企业效益不受影响。Based on the clear business problems, the paper screens effective input data and target variables, does correlation analysis between the parameters of the input variables, input variables and target variables, to select effective parameters. On the basis of data, it uses a neural network to establish forecasting model, and gives detection results, evaluates the model by operating the data in the actual business. Through the model to predicate of the loss of customers, and give warning signals, so that the enterprise can make business decisions, keep related users, ensure enterprise efficiency is not affected.

关 键 词：数据挖掘 相关性分析 方差分析 数据建模 神经网络 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]