A Research Of The Evaluation Of The Value Of Retaining Churning Customers Based On Neural Network

With the advancement of science and technology, as well as the increase of firms, thecompetition for customers among telecom operators gets fiercer and fiercer. As the number oftelecom operators customers can choose is becoming larger, the operators are faced with seriousproblems of consumer loss. Not all the losing customers are valuable to telecom operators,therefore, telecom operators have to judge which customers are valuable in order to makeappropriate detain-strategies to retain these customers. This dissertation is aimed at building amodel of evaluating detain-value of the lost customers.The dissertation made a thorough research of decision-tree and neutral networking, and thendiscussed the model of detain-value of lost customers in wireless communication using thetechnology of data mining, with the help of SAS9.1. The dissertation includes the design of themodel, the construction and revision of the model. Finally, we use the data from one mobilecorporation in Guang Dong as our sample data to compute the detain-value of the lost customers.The result is very good, which means our model is performing well in estimating detain-value oflost customers. Based on the computation results, we made certain kinds of detain-strategies, tohelp telecom operators make wise decisions.The dissertation mainly has three conclusions. First, we have calculated probability of lostfor different groups of customers, and gave the corresponding customer characteristics, such asaverage communication fees and customer types; secondly, we separated the customers into threedegrees of retain-value based on our estimated retain-values. Also, we gave the correspondingcustomer characteristics; finally, we provided certain retaining strategies for customers withdifferent degrees of retaining values, which would be of very practical use for mobilecorporations.