The Association Analysis In Mobile Communication Station Alarm And Design Of Station Remote Supervision System

With the increasing demand for communication service, the expansion of the size of the network, the increasing complexity of network structure,leading to more and more base station network fault alarm. A large number of irrelevant alarm flooded the real fault alarm, resulting in difficult to find, locate, and removal the fault alarm. In order to ensure low fault operation of the network, we must have effective monitoring and maintenance management for base station alarm, so the fault can be timely removed and reasonable forecasted. Alarm correlation analysis is an important technology of the alarm management, through correlation analysis to the massive alarm we can got the fault root reflected alarm, it can improve the maintenance staff's working efficiency, and ensure the high quality of the network operation.This paper has taken part of the alarm data from actual base station alarm database.After analysis we got the characteristics of the base station alarm data. Then,in-depth study on association rules mining algorithm based on the theory of data mining, the new AprioriTid algorithm improved based on the AprioriTid algorithm is used in the base station alarm association rule mining. we got the meaningful rules through post-processing. The maintenance staff can use these rules to filter,locate and predict alarm.Also,these rules can help the maintenance staff a better understanding of network operating conditions. This paper mainly do the following aspects:(1) Design the station inspection's automatic management process, the base station access warning process and flash alarm handling process.According to the above processes designed base station remote supervision system.(2) In-depth study on AprioriTid algorithm, and Combined with the application needs improve the algorithm named TP_AprioriTid algorithm based on multi-threaded, database partitioning, transaction and project compressed.(3) Set up a base station alarm association rules mining model and the improved algorithm applied into alarm association rules mining, through post-processing we got the meaningful rules.