Analyzing And Modeling Of Telecom Customer Data Based On Customer Segmentation

As the development of mobile communication, it is very important to serve the users. Nowadays, getting and evaluating User Experience is not only the key task for network operators, but the key work for different kinds of businesses as well.However, as the number of customer is huge, it is hard to get User Experience. Furthermore, because of the subjectivity of User Experience, the data of User Experience have deviation. It makes the User Experience Analysis the hotspot of research. The paper work on User Experience Analysis, aim to build the mapping between network data and User Experience data, which can be used to evaluate the Quality of User Experience by using Measurement Report. The work is listed as follow:(1) Data getting, analyzing, and preprocessing.(2) Proposing a new mapping of KPI-KQI-QoE which is able to evaluate User Experience through network data. The mapping is consist of two part, one is the mapping between KPI and KQI, and the other is the mapping between KQI and QoE. The paper works on the mapping between KPI and KQI, and proposes a method of analyzing user experience, which based on customer segmentation. The method can evaluate User Experience by historical data.(3)Developing a system of analyzing User Experience by using our research.(4) Experimenting and analyzing using real network dataIn addition, facing a mass amount of data, we propose new methods on customer segmentation and modeling.(1) In customer segmentation, we propose the adjustable multi-pass clustering algorithm. The results of the experiments show that the algorithm is effective to customer segmentation on telecommunication.(2) Facing the huge number of telecommunication data, a cooperative modeling method based on improved Support Vector Machine is proposed, which not only improve the ability of computing by parallel computing, but raise the precision by weight average method.