Font Size: a A A

Customer Churn's Analysis And Application Of Telecom Based On Data Mining

Posted on:2006-05-04Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q ShenFull Text:PDF
GTID:2168360155972391Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Nowadays, with the reform of domestic telecom, every operation companies have spread drastic competition in many areas, such as big customers of enterprise, distance operation, IP operation, mobile operation, and so on. Tradional market stratagem, which is focus on technique and production sell, has been replaced by new tactic, which concentrates on customer and service. It means that customer resource has become the key of enterprises' competition. Furious competition leads to the unstable state of customer, and operation companies almost all confront with such a serious problem--customer churn. Just for a great quantity of customers' vanishment, it has brought great loss to many companies. In such a severe circumstance, it has become one of the focuses of operation corporations that how to avoid customer churn and carry out retainment. Then managers find it important for operation enterprises in telecom to predict customer churn based on customer character and customer behavior, and make effective measures to realize customer retainment and customer attainment, which is also one of the most important projects on customer research and consultation in telecom. Author discusses customer value primarily first, and applies it to customer churn. At the same time, according to customers' historical data, this paper applies many technique of data mining to the research of customer churn, such as RBF NN, clustering, etc. As for those problems exist in corresponding research at present, author gives a detailed scheme to solve them, such as customer profile, attribute contraction, churn reason analysis, churn prediction and control tactics, and so on. At last, this paper succeeds to establish customer churn prediction model, gets the churn probability of each customer, and produces a possible churn customers' list. Through applying it to practice, author rectifies the effectiveness of customer churn prediction model, and gets a satisfied result, which can be used to instruct the actualization of market tactics in telecom.
Keywords/Search Tags:Data Mining, Customer Relationship Management, RBF NN, Clustering, Time Serial
PDF Full Text Request
Related items