An evolving process mining algorithm based on stable stages analysis

The Purpose of process mining is to discover the actual process of the running system by analyzing its runtime log information. Traditional process mining algorithms are based on hypothesis that the process to be mined is unchangeable, so they can´t mine evolving processes correctly. This paper proposes the concept of stable stages analysis, the SP-A stable stage analysis algorithm, and an algorithm based on stable stages analysis for mining evolving processes. It recognize stable stages based on process log, and then continuously mine the process logs of each stage, finding process model according to each stable stage, and to identify the evolving history of evolving processes. By experimental evaluations, the validity of the algorithms is proved.