Genetic Programming Is Applied Research In The Mining Of Non-complex Business Processes

Process mining is the automated acquisition of process models from event workflow logs.Although many process mining techniques have been developed, most of them have problems when mining processes that contain special constructs and/or when dealing with the presence of noise in the logs.Moreover, most of them focus on mining models from the behavior prospective shown in logs while ignoring the structural perspective of the mined model, which directly impacts the readability and quality of the model.To address these problems, we've proposed a genetic programming approach to mine a simplified (i.e., uncomplex) process model.Genetic programming (GP) is robust to noise and it breakthrough the express of complex structure in process models using tree-based individual representation.In addition, the "uncomplex" fitness derived from process complexity metric provides a guideline for process model to a simplifier one.The GP-based uncomplex process mining approach is proposed in this paper. First, the problems in current process mining techniques are shown to emphasis the need for mining uncomplex process model.Second, the individual representation in GP (called ECM) is defined to mine process models.In order to mine uncomplex process, the key issue of how to combining the process complexity with the GP-based process mining method is discussed. And it is implemented by defining the "uncomplex" fitness derived from the structuredness metric(SM),which has been extended to the ECM version. At last, experiments have also been conducted to evaluate its effects.