A Study On Applications Of Data Mining Technology In Network Alarm

Fault diagnosis and localization is the vital core of the network fault management. When the faults take place in the networks, it is necessary to find the locations and the causations of the faults accurately as soon as possible in order to get rid of the faults and recover the networks' function in time. The alarm correlation analysis is an important approach of fault diagnosis; it plays a crucial role in network fault management. However, it is difficult to acquire necessary knowledge to build an alarm correlation system of specific network because of its complexity. Data mining is a new means to find the knowledge of alarm information. Data mining provides a new approach of the knowledge updating during the alarm correlation analyzing.There are lots of alarm data in telecommunication alarm database. A lot of usable information is contained in these data, which reveals the status of telecommunication network. Alarm data is very suitable for alarm correlation application and the result of application can be used to network to improve the fault management of telecommunication network.According to this situation, this paper proposes to use data mining technology for the analysis of alarm data. The main works and contributions in this paper are as follows:(1) The article firstly analyzes the telecommunication alarm, and presents the aim of data mining technology in application of alarm correlation, and then gets the conclusion that it is feasible to use DM technology to mine the alarm association rules.(2) Secondly, this article introduces the conception of data mining and Alarm mining briefly, illustrates association rule mining algorithm in detail. While Apriori algorithm is used for data mining on alarm data, it will generate large number of candidate frequent itemsets. To avoid this weakness, we describe FP-Growth algorithm in association rule discovery in alarm data. In addition, aiming at the actual alarm data, the article implements a contrast experiment between Apriori algorithm and FP-growth Apriori.(3) By analyzing the property of the alarm data, the paper regards an alarm pattern in a time window as a sequence pattern. Based on the above fact, a frequent alarm sequential pattern mining algorithm named FASPMiner is presented. This algorithm adopts a prefix-projected pattern growth approach to decompose the task of mining original alarm database into a series of smaller tasks of mining locally projected alarm database with a 'divide-and-conquer' strategy so as to mine the frequent alarm sequence patterns contained in the alarm database finally.The extensive performance analysis show that suggested approaches in this dissertation are efficient and are competent for respective task of mining frequent alarm patterns.