基于压缩矩阵的通信网告警关联规则挖掘算法

Algorithm of mining alarm association rules based on compressed matrix in communication networks

提出了一种基于压缩矩阵运算的电信告警关联规则挖掘算法。它解决了apriori等算法需多次扫描数据库的问题,通过扫描告警事务库并进行压缩变换得到压缩告警关联矩阵,对关联矩阵进行运算得到告警间的关联规则。仿真实验证明,该算法与apriori等算法相比,时间效率有了明显提高,同时有效节约了存储空间。

This paper introduced an algorithm of mining association rules based on compressed matrix. Compared with the apriori and other apriori-based algorithms, which visiting database for many times, the algorithm visited database only once. By scanning the database and applying a compression technique at the same time, the algorithm converted the alarm database into compressed matrix, and then calculated the matrix to get the association rules. The performance testing of the algorithm indicates that compared with apriori algorithm, this one need less memory and has higher temporal efficiency.