Process mining aims at discovering process models from event logs,in complex information systems in the process of people’s subjective understanding of the business model of the process model and the actual deviation,by means of process mining technology can objectively log data generated from the system data of the more accurate process model.However,due to the complexity of the noise and the process model in the event log,some process mining methods can not do the optimization of the process model.Genetic process mining is a kind of process mining method,which not only can dig out all the common structure of the process model,but also has high robustness.But the algorithm may need several iterations to search the most ideal model,the calculation of the degree of adaptation in each iteration process is larger,resulting in the overall efficiency of the algorithm is not high.In order to improve the efficiency of genetic process mining,this paper presents a series of improvement measures for the efficiency of genetic algorithm.First,according to the searching efficiency of genetic operators,proposed a causal matrix with weights,and the basic crossover and mutation operators are improved;second,the genetic process mining algorithm based on log sampling,improves the genetic process mining face of handling large amounts of redundant log,reduces the computation log replay the scale of genetic adaptation process;third,in order to improve the fitness calculation efficiency,puts forward the concept model of BPST,and on this basis to provide a model segmentation method and parallel fitness calculation method.In this paper,the experimental results show that the improved genetic algorithm has higher efficiency,and the basic idea of the algorithm and the feasibility of the algorithm are verified by the experimental results. |