关键词: 分层过程挖掘, 分层Petri网, 生命周期事件日志, 质量评估

Abstract: Existing process mining techniques usually support discovery of flat process models from event logs,which are not able to investigate the nesting relation among activities and hierarchical process model mining.For this problem,a novel approach to automatically discover hierarchical business process models was represented as hierarchical Petri nets from event logs.To handle noise and infrequent behavior,the notion of nesting ratio was introduced to quantify the probability of nesting,and a parameter was introduced to allow users to set a desired level of nesting threshold in a flexible manner.The proposed approach had been implemented in the open-source process mining toolkit ProM.Using both synthetic and real-life event logs,several existing state-of-the-art process discovery techniques was compared to our approach,and the results showed that the proposed approach was superior to these exiting approaches in discovering hierarchical business process models.

Key words: hierarchical process mining, hierarchical Petri nets, lifecycle event log, quality evaluation