Research And Application Of Data Warehouse System In Telecommunication Region

Along with the breaking of the monopolize pattern, the competition among the domestic telecommunication operators is fiercer and fiercer. In order to get the vantage, it is very necessary for telecommunication operators to build a powerful solution, which can convert the mass telecommunication data into reliable business tactic information, and supply powerful ability in market analysis.Of late years, telecommunication operators got a mass of data stored in the relational database. However the traditional rational database cannot be used to find hidden information from the raw data. How to analyze historical data and get knowledge has become a factor for telecommunication operators to win the competition Data warehouse and Data mining technologies provide an effective way to solve this problem.Data warehouse is a kind of new-type method to organize and store data. It organizes transaction data that enterprises accumulate again with multidimensional data modeling and sets up new data set, which is specially used for supporting data analysis and making policy for enterprises. Data mining is the strongest data analysis means in the warehouse field at present. Its analysis method is utilizing the known data to find out the implicit business rules through setting up mathematics model.At first, this paper has explained the background of the project, the application trend of data warehouse, the purpose and content of the research. Second, the technology of data warehouse including the concept, the characters and the critical technology of data warehouse and data mining is analysed. Third, it also analyses the data warehouse's application condition and researchs the data mining's concrete realization in telecommunication operators.One of the important research works in this paper is designing a loss of customers' prediction system. It discusses the application that the decision tree is used in the telecommunication marketing field. It is building customer losing model using Data Mining method and decision tree arithmetic that mainly according to the character and the consuming behavior of the customer losing or not in the telecommunication enterprise. Marketing policies are also established according to customer losing model. In this portion, we discuss the construction theory of the decision tree and the shaving theory of the decision tree avoiding the excessive training. Then this paper introduces the application of clustering Methods in the telecommunication marketing field.By using the decision-tree to analyse the client losing, the forecasting model of the client losing is set up. Clients that have the largest probilitiesto lose, and their correlation factors are analysed, the object is to provide the decision gist to market managers and decision-makers, who need to institute the policy to retain those clients.