Abstract:

Modern telecommunications networks are very powerful, and there is a sharp increase in business volume, and network is becoming large, complex, heterogeneous. In order to ensure the normal operation of the entire telecommunications network, we must improve the efficiency of processing network alarm, and identify the root causes of failure rapidly. Association rules algorithm of data mining can effectively analyze the relevance of things happened in the network warning system. It can also accurately locate the failure of a network application.We improved the traditional association rules algorithm corresponding the characteristics of alarm data. We improved mining efficiency and achieved a rapid diagnosis of network failures.

Key words: alarm diagnostic, data mining, association rules, Apriori algorithm