Research And Implementation Of Early Warning Forecast Method In Monitoring And Analysis System Of Mobile Data Service

At present, the market competition becomes increasingly intense, which forces telecom operators to provide new service to meet users' needs constantly and ensures that the telecom network can provide reliable, stable and high-quality services continuously. With the accumulation of time and the increment of scale and complexity of telecom network, the network performance data that is collected in operation shows exponential growth. This data can best describe the quality of network performance and network optimization behavior.Effective management and analysis of this data will be a great significance of optimizing and allocating network resources. However, traditional network management system lacks of capabilities in integrated data management and data analysis. Therefore, telecom operators hope to manage and use this data unifiedly in network management system to upgrade the quality of service in telecom network managment.Monitoring and Analysis System of Mobile Data Service starts with mobile data service that is a kind of updating frequently service and enhances the data analysis capability of network management system. In addtion, this new system improves the function of preventing faults in advance through inter-transactional association analysis method of data mining.This paper mainly carries on following several works:(1) Participate in architecture analysis of Monitoring and Analysis System of Mobile Data Service. Mainly responsible for the design and implementation of early warning forecast subsystem.(2) Present an improved algorithm of inter-transactional association analysis named AFP-Growth in early warning forecast subsystem.(3) Design and implement AFP-Growth algorithm module, verify it in early warning forecast subsystem and achieve good result. This paper analyzes the solution of monitoring and analysis in mobile data service, discusses the structure features of early warning subsystem and introdueces the design and implementation of early warning subsystem. At last, the design and verification of the improved algorithm of early warning forecast method are discussed.