基于频繁项集挖掘的发布/订阅分布式系统运行模式识别

Running modes recognition of publish/subscribe distributed system based on frequent itemsets mining

为保障发布/订阅分布式系统的安全运行,需实时监控系统的运行模式,以发现系统的异常状态,因此识别出系统中正常的运行模式是状态监控的首要任务。基于发布/订阅分布式系统的运行特点,提出了一种加权频繁项集挖掘算法用于运行模式的识别。实验结果表明,该算法能有效挖掘出发布/订阅分布式系统中的运行模式,同时相较于Apriori算法和FP-growth算法有较好的性能。

In order to ensure the safe operation of the publish/subscribe distributed system,it is necessary to monitor the running mode of the system in real time to discover the abnormal state of the system,so identifying the normal operation modes of the system is the primary task of the state monitoring.Based on the running characteristics of publish/subscribe distributed system,a weighted frequent itemsets mining algorithm was proposed for running modes recognition.The experimental result shows that the algorithm can effectively mine the running modes of publish/subscribe distributed system,and has better performance than Apriori algorithm and FP-growth algorithm.