The Analysis And Prediction Of Customer's Loyalty In Telecom Industry

As the gradual development of telecom market and the competition position beingseverity, many telecommunications companies are face with the problem of customer'slow loyalty. To survive or maintain an advantage in such a competitive marketplace,many telecommunications companies are trying their best to provide many kinds ofhigh-costing promotion activity for attracting customers. In virtue of investigation, thecost of company's attracting a new customer is five to ten times of company'smaintaining an old customer. Thereby, how to maintain customers of high loyalty andstability customers of low loyalty is the urgent problem of many telecommunicationscompanies. In this thesis, we analysis each factors influencing customer's loyalty, andset up a model to predict customer's loyalty by using C4.5 decision tree. These can beprovided to telecommunications companies to solve the problem of classifyingcustomer's loyalty.Constrained by limited customer profiles and general demographics such as gender,age and occupation, the proposed approach work over the customer's loyaltyclassification based on the telecommunications companies'existing data records. Byquestionnaire of a few customers, we educe the index of their sensibility loyaltyincluding. By picking up the characteristic variable from the call records and customer'sbasic data, we educe the index of their behavior loyalty. By analytic hierarchy process,we can calculate the sample customer's loyalty. We apply C4.5 decision tree to learnfrom the sample customer's loyalty and get the loyalty classification of many customers.We use C4.5 decision tree induction technique and boosting computing to set up aclassification model for customer's loyalty predication. The results showed that theproposed technique was effective in the customer's loyalty classification and it can beapplied in reality.