The Construction And Application Of CRBT Potential Customer Model Based On Decision Tree

CRBT has been welcomed by the majority of users as it was introduced, it hasbeen one of the most growing value-added services and the main source of newrevenue in the telecommunication industry. After several years of rapid growth,CRBT is now nearly mature stage. But the characteristics has emerged which includeslowness of incremental market development, increasingly saturation of existingmarket and slowly increased penetration. In early days, because of lack of accurateidentification of target groups, CRBT promotion to popular customers not onlywasted sources but also planted hidden dangers for customers purity, leading to hignrate of accounts-canceling and low activities. The main marketing problems inCRBT develop process are how to find the characteristics of the existing CRBTcustomers and to raise the penetration and activities of customers and to deduce therate of account-canceling.There have been some research results about the CRBT service market at homeand abroad, but most of them are confined to the platform of CRBT analysis system.in accordance to the existing situation that telecommunication market possesrelatively complete data about consumer behavior characteristic, this article proposesa CRBT potential customer model construction method based on Decision TreeClassification which is contributed to locate CRBT potential customer, and apply it toguide CRBT to orientate their market position and promotion strategy.Firstly this article describes the background and significance of this subjectinvolving the develop process and overall status of our country’s telecommunicationCRBT.; Then we explain and discuss the basic knowledge about the Decision treeclassification used in our model in chapter two; after that we analyze the functionaryrequirements and data requirements of model construction in chapter three, and makesome improvement of basic decision tree classification algorithm for modeling data’scharacteristics; There is detailed description of model construction process. Heregives an example about one mobile company’s actual data. By researching the usecharacteristics of new CRBT customers, we analyze customers’ behavior attributes;By using the decision tree technology of DM and process CRSIP-D which followscrossing-fields DM standards, we offer a complete proposal to solve the problemabout CRBT potential customer model construction, we also discuss and analyze theimplementation key points of every step of the model construction process whichincludes “commercial understanding-data preparation-model evaluation-modelissuance”, and test the practical functionaries of the model with actual data. In the last chapter, we make more detailed description about maintenance and update of themodel.We have tested that this model can greatly dig out the potential customers ofCRBT so that it can locate the target customers precisely, seek more new customersefficiently, reduce the rate of account-canceling. All these provide effectivetechnology support for adjusting the business strategy and promotion timely andcarrying out the fine marketing activities. At the same time, it is of great significancein theory research and engineering practice.