Constructing Customer Leaving Analysis Model Based On Data Mining

In recent years, as the competition among Communication enterprises becomes more and more fierce, customer losing has been a serious issue that puzzles mobile communication enterprises all over the world. As one of the largest subsidiary companies of China Mobile (HK) Limited in Mainland, ZMCC expanded the business rapidly year by year and the data of ZMCC is increasing at a geometric series rate. The basic functions of the operational analysis system in ZMCC are still collecting data, getting statistics, displaying results. However, the problem of customer leaving is related to a large-scale set of customer data. If we still make use of superficial information of data by traditional simple statistics without mining deep information among data, it's difficult to discover complex rules about customer leaving out of large-scale data.As a powerful technology, Data Mining can help enterprise to discover the most important information from data warehouse. Some software companies have provided several general tools, which can't deal with the dynamic requirements of specific companies like ZMCC. We are intended to take research on this issue and apply data mining technologies to construct customer leaving model. The main contributions of this thesis are as follows: analyze and research the application of data mining in mobile customer leaving analysis; research and implement the closed loop model for leaving control; demonstrate the scenario and outcome of the leaving predictive model;...