Design And Implement Of Off-grid And Arrear Prewarning System For Telecom Key Customers

China’s telecommunications industry is facing fierce inter-industry competitions and the business model of telecom providers has been gradually converted from the "technology-driven" mode to the "market-driven" and "customer-driven" mode. The concept of "insight marketing" and "precision marketing" through in-depth data analysis based on information, value and behavior of customers, has been widely accepted by the major providers. They have also been seeking to save the marketing costs and improve the marketing efficiency through mining data from various application systems.With respect to these circumstances, telecommunications enterprises are seeking to build intelligent business platform to save marketing costs and improve marketing efficiency. Compared with other soluations, building a data mining platform is an more effective way for the intelligent platform construction. The traditional data mining technology applied in the telecommunications industry has the limitations such as long developing cycle and deficiency of flexibility, etc. With the development of the SOA technology, a data mining technology application platform based on Web Services has also been brought up, which makes it feasible to construct integrated, flexible and customizable telecommunications data mining platform.In this paper, regarding potential risks brought by large telecom customers, We jointly consider the the status of business sale and mangement and the needs of business marketing on China’s telecommunications providers, use the data prewarning and data mining techonogies, and further propose a new SOA model which focuses on the departure of key customers and the prewarning of arrear. Specifically we give the requirement analysis, general design, the detailed design and the implementation of the model. In this model, we apply the risk prediction technology into the data analysis of telecommunications industry, analyze and predict the consuming features of customers, derive the probability of the departure and contract violation of customers, and provide objective and reliable data support for the establishment of early warning mechanisms.