The Churn Prediction System Of Wireline Telecom Based On Data Mining

Chinese telecommunication industry has been further opened with joining WTO. Internationalized market circumstances request the telecommunication operation company that the domestic management of telecommunication catches up international advanced level rapidly in order to face international competition of telecommunication industry; meanwhile, with uninterrupted reform and recombination of domestic telecommunication, the market circumstance is going to become reasonable gradually, competition is becoming intensive day by day. The competition status of domestic and international telecommunication industry request that running concept and management pattern of Chinese telecommunication industry have a higher leap, which will center on the production, on the basis of business windows, and gradually transform into on the center of customers, data and information. Especially for the fixed telephone network operation company which the author is working for, the fixed telephone network market is basically in saturation, incremental market is shrunk gradually, so it is very important how to keep current market, analysis and prediction of customer churn will be the most important point in customer analysis.The purpose of this article is to establish an appropriate prediction model of fixed telephone customer churn on the basis of analyzing the characteristics of fixed telephone customer churn and combining the support system of fixed telephone network operation, to provide the supports for customer maintaining and staying in the current stage, to provide the beneficial lessens for reference for more subjects after the completion of management analysis system.Data warehouse and data mining technology are the basis of customer churn prediction system. Data warehouse is a data collection which is subject-oriented, integrated, non-volatile and time variant and used for supporting the management decision-making. A data warehouse consists of four parts: data source, data memory and management, OLAP server and front-end tool. A data warehouse includes (or comes from) the data groups of several departments. Data mining is a process of picking up some unknown and potential useful information and knowledge from a large amount, incomplete, noisy, indistinct and random data. Common data mining algorithm includes: classification, estimation, prediction, relativity grouping and correlation rules, clustering, etc. Data mining has a standardized process.