Research On Application Of Data Mining Technology In The Customer Maintenance Work

In this full business operation stage of communications industry, the market competition is becoming increasingly fierce, the competition level between operators has been transfer from the traditional network and product to service capability of customer maintenance. Leading in service has become the development strategy of the major telecommunications operators. Operators are constantly strengthen quality control, improve customer perception, so as to improve the customer retention capacity and the stability of user.With the development of data mining technology, variety of data mining algorithms are derived, including classification and clustering, regression and association rule, variation and deviation analysis, the Web page retrieval and mining. These methods, respectively, to mine data from different dimensions. This paper focuses on the’classification’methods of data mining. Classification can be applied to the customer classification, customer analysis of attributes and characteristics, customer satisfaction analysis, churn prediction and so forth.This paper concerns the business characteristics and the data of telecommunication company (Ningde branch company of China Telecom), research on application of Naive Bayes methods, after compare with many other data mining algorithms. This paper improve the depth and accuracy of customer maintenance, by introducing Naive Bayes method, which is enhanced by improving the existing discretization algorithm, the ChiMerge, which makes the classifier more sensitive about high loosing risk customers, and perform a better classification effect on high risk population.By introducing data mining methods, massive dataset about customer telephone fare, package and fusion rate is analyzed effectively. With respect to the mining result, operators can carry on targeted customer care and maintenance, improve the quality and stability of the network users. With the development of data mining method, customer maintenance work will definitely better improved by this technology, which make customer maintenance more intelligent.