Research On Online Conformance Checking Method Based Of Business Process On Log Repair

As business process management systems continue to evolve,organizations need tools to improve the transparency of their internal processes.The process discovery algorithm can be used to mine the relevant process model from the event log,but the ordinary process discovery algorithm cannot mine the model that fully conforms to the actual situation of the process model.The starting point of process discovery algorithm is event log,so the quality of event log affects the model quality of process discovery algorithm output.At present,many scholars realize that the accuracy of the output model of process mining algorithm can be improved by repairing event logs.In the past,the method of simply and directly deleting uncommon or inconsistent instances of process running is no longer applicable to the huge database.Therefore,it is very necessary to use the internal relationship of business processes to identify and repair event logs.With the continuous development of the society,the business process management system becomes more and more complex.The event logs and various data that record the operation of the business process are increasing explodes,which makes it impossible for the organization to store completely and handle it offline.A posteriori conformance checking method has been unable to meet the ever-developing business process management system.At the same time,the common prefix alignment technology can’t calculate the conformance value of the running business process,and can’t immediately alert the organization manager when the problem occurs,so the development of online conformance checking technology is particularly important.This paper mainly uses the relevant technology of process mining to process the event log,and uses the relevant theoretical knowledge of Petri net and behavior profile to build the initial target model,and then optimizes the model.In order to improve the conformity of the target model,an event log repair method is proposed based on the relationship between the subsequences of the service flow.Finally,in order to detect the biased part at an early stage and improve the effect of conformance checking,an online conformance checking based on hidden Markov model is proposed by using the special structure of hidden Markov model,and the proposed log repair method is evaluated.Specific research work is as follows:(1)Aiming at the problems existing in the parking reservation management system,a process mining algorithm based on business subprocess is proposed,and then a complete Petri net is constructed by analyzing the successive relationship between each business subprocess.The log of parking reservation process is used to optimize the original model,which improves the practicability and user experience of the system.(2)Aiming at event logs with outliers and incomplete traces,an event log repair method was proposed to improve the quality of event logs and provide a more accurate starting point for process mining.The event logs are identified,incomplete traces are screened out,and the possible problematic subsequences are replaced by using the context and coverage probability of the subsequences,so that the possible outliers can be repaired and the quality of the event logs can be improved.(3)An online conformance checking method based on hidden Markov model is proposed.Using state estimation and probability function in hidden Markov model,on-line conformance checking is carried out.The online conformance checking is divided into two aspects: orientation and consistency.At the same time,the degree of conformance is divided during detection.In the offline stage,the coverage probability of the context relationship in the event log is calculated,which provides parameter estimation for the online conformance checking stage and improves the detection accuracy.Figure [33] Table [28] Reference [87]...