关键词: Petri网, 行为轮廓, 事件日志, 过程挖掘, 拟间接依赖关系

Abstract: Business process mining aims to dig out the process model to meet the needs of people from the event logs. Most of the previous methods are based on the direct dependencies between events to build process models. Previous methods have some limitations. This paper proposes a method of optimization analysis of process mining based on quasi indirect dependency. Firstly, according to the event log, the initial model is constructed on the basis of behavior profile. Then, this paper finds the transition pairs with quasi indirect dependency through the basic constraint body of the ILP-based process discovery algorithm based on example event log. The model is improved and the optimized model is mined. Finally, a concrete example analysis is given to illustrate the effectiveness of the proposed method.

Key words: Petri net, behavior profiles, event log, process mining, quasi indirect dependence