Process Mining Based on Statistic Ordering Relations of Events

Abstract

An explicit process model is vital in business processes. However, it is complicated and time consuming to create a workflow design. Also discords usually occur between the perceived management processes and the actual workflow processes. Under this condition, the process discovery techniques emerge. The aim is to rebuild a workflow model (e.g. a Petri net) of a business process based on the execution log. The model should give an abstract representation of the system and reproduce the log. This model can be further applied for process redesign/improvement and performance/reliability evaluation. In this paper, we present a new algorithm derived from α-algorithm for process discovery in term of Petri nets, where statistic long distance causal relationship is taken into consideration. Also this algorithm covers some shortages in α-algorithm.