Application Of Data Mining Technology In Network Management Of Telecommunication Networks

Telecommunication network are composed of many different subnets. Each of the subnets has already established its own network management system. With the development of information industry, the size of telecommunication network becomes larger and larger and the complexity of structure becomes greater and greater. It will become a usual matter that dozens of alarms of different devices exist on a same alarm platform. So, the related management will become more and more difficult. It will be a very critical problem to make a swift respond and accurate handling to the alarms. This will guarantee the free-flowing of the networks; promote the performance and utility ratio of the networks. Under the traditional network management, a mass of staff are needed, who must have deep knowledge of telecommunication network and experience in the maintenance of telecommunication network. On the other hand, using the artificial analysis, the maintenance staff will be tremendously misled if the partial alarms are not sent timely to the network monitoring platform. At the meantime, it can not adapt to the frequently change of network. So, the alarms data in the database of network management system need to be analyzed and processed inartificially.This problem can be solved by the data mining technology. The data mining is aimed to the alarm data in the database of network management system. The alarms will be filtered and located swiftly by studying the method of correlation analysis and the relevance between alarms.We analyze the features of alarms and handle the raw alarm data beforehand so that they can be used to mining rules. The correlation analysis model for alarm data is brought up. According to it, we enhance the association algorithm to make it suitable for alarm data and mine some available alarm rules. It is proved that data mining technology has a great potential to be applied in the management system of telecommunication networks.