Research On Process Mining Method Based On Incremental Log

Process-Aware Information Systems are widespread used in most contemporary organizations,which record a large number of business processes in the form of event logs,and process mining technology is used to extract useful process knowledge from the event log.It aims to find,monitor and enhance real business processes.The application of process mining technology has a positive effect to enhance the company's productivity,and it plays an important role in guiding the company's operation and saving operating costs.Therefore,process mining techniques are key technologies in business process management.The continuous improvement of information systems allows data in the business process to be sufficiently discovered and utilized,and process mining technology faces constant challenges.But existing process mining technologies are mostly based on static event logs,which are analyzed and studied on the business process stored in the database.In other words,for dynamic update logs,existing process mining techniques exhibit certain limitations,such as the newly generated log for a mining requires huge human material support,and may hide or even lose important models in the original model.Behavior,etc.Aiming at this problem,this dissertation proposes a process mining method based on incremental logs.First,an initial process model is mined based on the static event log,and then the difference between the incremental log and the original event log is analyzed and the initial process model is updated according to the difference.The updated process model can play back both static historical data and newly added event logs,which improves the suitability of the model,reduces the time overhead of process mining,and ensures that important behaviors of the original system will not be lost.The main tasks of this paper are:(1)Aiming at the problem of business process model change detection,a method to investigate the feasibility of model after model change from the perspective of sub-modules is proposed.The feasibility of the sub-module is investigated by calculating the correlation matrix and state equation of the sub-module.Combined with the actual changes of the company's business process,the method is verified,and the optimization sub-module is proved to be feasible through the simulation experiment.(2)Aiming at the problem of change and update of incremental log in business process,an optimization method of incremental log process model based on optimal approximate trace is proposed.First of all,on the premise of keeping the possible behavior relationships of all activities based on the log unchanged,the traces containing concurrent and cyclic behaviors are preprocessed,and then based on the idea of data structure string comparison,the optimal approximation between the incremental log and the original log is calculated Then,determine the change position of the optimal approximate track by comparing the difference between the incremental track and the optimal approximate track,and then find the corresponding change domain in the original model according to the optimal approximate track,and finally,according to different change types Adopt different optimization schemes respectively to obtain the optimized process model.(3)Aiming at the dynamic change problem of business process,based on the existing research,this paper adopts the behavior clustering algorithm based on trace similarity which changes the business process model according to the change of the process log provided by the user.Bind the user to the database and the company to the database.The user is only responsible for providing the process extraction log and accepting the business model,the database is only responsible for accepting and sorting the transmitted data,and the company is only responsible for analyzing the data and optimizing the model.The advantage of this method is to reduce the complexity of business process from the perspective of both sides.Figure[24]Table[10]Reference[99]...