摘要：

不完备日志挖掘由于日志中含有较少的迹,使活动间潜在关系不能被正确表示。为此,引入间接关系表示并提出一种针对块状并发过程模型的挖掘算法。根据活动在层次树中的位置,发现潜在的因果跟随关系并得到正确的因果跟随关系集合,在此基础上,挖掘出块状并发过程模型。实验结果表明,在处理并发完备日志时,相对α‖算法和IM算法,该算法具有较高的精确度与较低的复杂度,且挖掘结果更简洁。

关键词: 过程挖掘, 块状并发过程, Petri网, 不完备日志, 因果跟随关系

Abstract:

Incomplete logs mining can not represent the potential relationship between activities correctly because there are fewer traces in the logs.To solve this problem,an indirect relation representation is introduced and a mining algorithm for block concurrent process models is proposed.According to the location of the activity in the hierarchical tree,the potential causal follower relationship is found and the correct set of causal follower relationship is obtained,thus the block concurrent process model is excavated.Experimental results show that compared with α‖ algorithm and IM algorithm,this algorithm has higher accuracy,lower complexity and more concise mining results when dealing with concurrent complete logs.

Key words: process mining, block concurrent process, Petri net, incomplete logs, causal follower relationship

中图分类号: 

• TP391