Based On Data Mining Technology, Business Analysis System

With the pattern of monopolized is broken, the competition among the domestic telecommunication operators is becoming more and more vigorous, and the simple price competition causes the loss of all sides of the competition. In the situation that the difference in such respects as the service quality of the network ,etc. reduces gradually, in order to improve the market competitiveness, telecommunication enterprises are all seeking to improve customer's service quality method.They urgently need to improve the science decision ability, to strengthen the correct judgement in such aspects as market management.Customer resources are the most important key resource of operator, only the cognitive customer abundantly and carefully , understand the customer' s difference, could offer better services to customer , could improve the customer satisfaction and loyalty , bring the income and profit to operator, raise the market competitive position of the operator. The application that the data mining technology can help the operator to analyse customer's consumer behavior, discerning customer's characteristic, the auxiliary operator carries on the effective marketing and customer service.At present the mobile communication all have been build the buessiness and operation support system, accumulated magnanimity and electronical data, but each information systems are all each manage some of the customer' s data, and numerous customer data, action marketing datas, the accounts datas scatter in system of dissimilarity by the different data format and way of visiting, forming the numerous information detached islands, it is redundant and inconsistenting that the data exist in each information detached island, can't respond to the request that the data excavate in the course the data must have single view . At the same time the business systems are all the process systemof on-line affairs , the online affairs of real-time processing, can't adapt to the data and excavate and use the extensive , frequent search operation in the course .So we must set up customer information data warehouse of enterprise layer , gather different on-line affairs customer datum of process system, offer a datum environment to a correct , intact and single customer. At the same time the system must be able to also integrate the instrument to dig up of specialized data and data analysis , represent tools , offer the convenience that the policymaker uses and carries on performance analysis systematically.The thesis proposes to this state a mobile communication enterprise excavating technology on the basis of the datum deals in the analyticl system framework, combining enterprise's characteristic of mobile communication will analyse and design to the system, the reason fails to explain in detail the system realizes the course as space is limited. Mobile communication enterprises are big in data amount, choose the suitable data to excavate algorithms, improve network analysis efficiency, set up and manage one of the questions that the course of the analyticl system must be considered.The thesis aims at the key technique of the mobile communication undertaking management analysis system adoption, analysis the sum design procedure carried on the research, and combine the mobile communication undertaking characteristics logarithm according to scoop out the algorithm to make improvement according to the effective demand. In the course of data mining we strictly obey the CRISP- DM( the CRoss- Industry Standard Process for Data Mining) process model. Combine with the mobile communication's character of having numerous of data we use decision tree analysis method to carry on the customer' s loss analysis. During the customer subdivide procedure, we need to carry on the factor analysis because of the variable that the customer' s actionsinvolve numerous and strong relativity, will affect the substitution of the cluster algorithm, so we carry on the improvement to the cluster algorithm, before carry on cluster.We work hard to use fewer random variables to describe many variables which a kind of basic structure, thus decline the data dimension to a fit horizontal, thus raised the velocity of the cluster analysis.Finally , the thesis researches application of Data Mining technology in the customer' loss and subdivision. The Data Mining results accords with the customer loss rules accord with the production and decision law basically .which have certain meaning of guide to the analytical forecasting and also have reference value to the other aspects of customer research.