Font Size: a A A

Applications Of Data Mining In Mobile Users Recognition Of Telecom

Posted on:2014-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:J MaoFull Text:PDF
GTID:2268330428457354Subject:Applied statistics
Abstract/Summary:PDF Full Text Request
As the development of the communication and the popularity of mobile phones, the communication of our country is becoming more and more completive. The philosophy of the telecommunication has changed from original product-centric to customer-centric. The communication to improve their business by expanding the mobile phone uses has been very difficult. Currently, the most important carriers for the communication are to retain customers, especially high-value users and avoid customers churning.First, the project studies how to identify high-value customers of telecommunication. According to the understanding about high-value customers of staff of telecom, then the project identify using data mining methods to determines which users should high-value customers and analyzes the characteristics of high-value customers should met. In the analysis of high-value customers, the project has studied from two perspectives:the business perspective and statistical perspective. Finally, the results of the two methods are used to determine the final high-value customers. Secondly, the project studies how to identify the customers who will churn. As the competition of telecom is becoming intensive completive, no one telecom can sure his customers no longer will not turn to other carriers. Therefore, to predict which customers will be lost is the foundation for the carriers to retain his customers. Based on the understanding of business, the project first to define the churn customers and then using the algorithm of decision tree to analyze. Finally, the project compares the result of the algorithm:C&R, QUEST, C5.0and C5.0based on boosting, from which to choose C5.0as the final model.Finally, selecting high-value customers from the customers who were predicted have high probability to loss as the important customers for telecom to retain. Then the project analyzes retention measures of high-value customers who are predicted to be loss.
Keywords/Search Tags:Data Mining, High-value Users, Customer Identification, Churn Prediction, Clustering, Decision Tree
PDF Full Text Request
Related items