User Data Analysis Based On The Miss-call Protection System

With the development of the telecommunication industry, the competition between operators is more and more fierce, at the same time, the development of mobile communications also affect the fixed-line operators, so the fixed-line operators begin to use a variety of measures to attract users, they hoped to achieve the purpose of increase income, but the only way that can achieve this goal is to provide satisfactory service. The miss-call protection system is a value-added business systems aimed at reducing the number of fixed-line users missing call. It puts the voice access to miss-call protection system, and provides users with a friendly voice interaction. It helps users to continue call or send SMS, etc. However, the business is different from the previous missing-call SMS reminder, so which users should the business be open to, what are the different strategies for different users, this is the purpose of the thesis.In this thesis, we study the miss-call protection system, including the business background, system structure, business process, etc. Miss-call protection system is mainly for the enterprise users, so the main problem is about the selection of users. This thesis uses the clustering analysis of data mining technology to analysis the user data, and classifying the user. Then defined each type of user, in order to operators can provide the better personalized service. In this paper, the main work is as follows:(1) Design the structure of the miss-call protection system. Study the main problem of miss-call protection system: the analysis of user data.(2) Study the data miming knowledge, and insight into the miming method, especially the K-means algorithm. Analysis the disadvantages of the algorithm.(3) Using the SPSS Clementine software, with the improved K-means algorithm for building user data model of the miss-call protection system.(4) Show the classification process and results of the miss-call protection system user data. Then make definition and analysis of the results. Finally suggestions of the market strategy are put forward according to the different class user.