Research And Application Of Loss Of Customers In Telecommunication Based On Multi-Agent Data Mining Technology

Following the opening and growing-up of the telecommunication industryand the coming of the 3G, the competition among the telecommunication companies are getting critical. How to keep the customers is the core issue of the telecommunication companies.But traditional data mining system has a series of problems such as low efficiency. To overcome these deficiencies, it is necessary to propose a kind of new data mining mechanism.Agent technology is applied widely in many fields. Multi-Agent coorperation enhances the ablity of multi-Agent system to solve problem and makes the system more flexible. It would be a helpful method to use Agent technology to solve the problems in prediction of churn in telecommunication data mining.Firstly, the thesis introduces data mining, multi-Agent and analyse the deficiencies of the traditional data mining. The thesis introduces the traits of Agent technology and analyse the advantages of Agent technology in data mining. Secondly, an model of data mining system based on multi-Agent system was bulit up. the responsiblities and behaviors of each Agent in the model has been discussed. Finally, the model of data mining system based on multi-Agent system has been applied to the Very Important System. Predictive model about churned clients has been presented and detailed ways to establish model has been provided.Agent technology is applied prediction of churn in telecommunication, which can save network bandwidth and overcome limitations of resources, space-time distribution and function in the traditional data mining so that it can achieve the system objectives better.