Process models can be discovered from event logs generated by the enterprise information system. As business processes' frequently changing, some activities in event logs may belong to different choice branches, while the actual model can only replay activities on the same choice branch. Thus, the model needs to be repaired to describe the business process accurately. For activities occurring in a choice structure, although the model repaired by the existing methods has good fitness, its structure may have changed a lot. This paper proposes a new model repair method based on logic Petri nets. First, we define a new event log type-choice deviation sub-log. Then, we find deviation positions according to the principle of token replaying. Next, we add bridges among choice branches to repair the model. We conduct experiments on some cases from cancer treatment processes in a hospital. The effectiveness and correctness of our method can be illustrated. The model repaired by our method is similar to the original one, and it has a higher precision compared with the existing model repair methods.INDEX TERMS Logic Petri net, process mining, model repair, choice structures.
Models of business processes can be discovered and improved by process mining techniques according to event logs generated by enterprise information systems. With the changing of business processes, some new activities that have choice relations with original activities appear in event logs, while the original model cannot replay them. Thus, a new process mining technique named model repair is developed. However, existing model repair methods cannot accurately find the positions where to add these new activities. This paper proposes a new model repair method based on logic Petri nets, which can add new activities as choice branches to the original model. First, two order sets for an event log and an original model are constructed by redefining order relations. Then the deviations related to new activities are collected. Next, the model is repaired by adding new activities as choice branches or constructing new choice structures. The correctness and effectiveness of the proposed method is illustrated by some case studies and experiments. The precision and simplicity of the repaired model is improved by our method comparing with other methods. INDEX TERMS Logic petri net, process mining, model repair, order relations, choice structures.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.