Along with internationalized expansion of telecommunication industry, the whole industry is quickly evolving into market economy from primary planned economy. Customer losses in shifting teleservice become less and less since marketization of telecommunication become more and more sufficient, telesevice standard closer and closer, and differentiation smaller and smaller. Operators are starting to shift key point to enhancing customers'loyalty and how to maintain and cater customers has, at present stage, become an important problem they must to tackle.Data mining is a process during which relationship between model and data is found among vast data resources by way of various analysis tools. As a typical data-intensive industry, telecommunications own large numbers of business data implied various useful information. the paper puts emphasis on design of data warehouse and selection of data mining model so as to integrate with Tietong's practical situation, aim at the demand of customer churn control among Tietong's landline users and center on application of data mining technology in customer churn control. The paper adopts CRISP-DM method, commonly used within the industry, to control entire data mining flow and chooses CHAID calculation method of decision tree to set up Tietong's customer churn control model.Through various applications of data mining technology, customer churn control based on thus technology can help to find connotative information on lost customers, foresee in advance those who have the possibility of churn and stabilize market share. In the meantime, it is hoped that Tietong can be in a position to be enlightened in some extent when carrying out customer retention strategy. |