Research On Business Process Mining Based On Communication Behavioral Profile In Petri Net

With the arrival of the era of big data,various event data are recorded in information systems,and the number of event data is huge and it is difficult to handle.For such problems,process mining technology can better achieve the link between the event data and the process model,and related operations on the event data.In order to mine the process model applicable to the market demand from the extracted event logs,certain infrequent actions in the event log need to be optimized.Mining and optimization under infrequent behavior of business processes has become an important hot research topic.The current study of the infrequent behavior of business processes focuses on the removal of infrequent behavior as noise.This approach ignores the importance of part of the behavior and makes the mining business process model lose some of its practical significance.The problems of infrequent behaviors of the process mainly include the mining of quasi indirect dependence,the mining of hidden transition in business processes,and the mining and optimization of conditional infrequent behavior.In the mining process model,if only the direct dependencies between events in the event log are considered,the excavated model can hardly meet the actual requirements.The existing methods have certain limitations in digging the accuracy of hidden transitions.For the mining optimization of the infrequent behavior of business processes,existing methods seldom analyze or ignore the communication behavior between modules from the perspective of data flow.Based on the above process infrequent behavior mining problem,a corresponding algorithm is proposed.The main literature of this article is expressed in the following aspects.(1)With regard to the indirect dependence between event events,a process mining optimization analysis based on quasi-indirect dependencies is proposed.Based on behavioral contour theory,first,an initial model is constructed based on the event log.Then,under the execution log,by finding the basic constraints of the algorithm based on the integer linear programming process,the transition pairs with quasi-indirect dependencies are searched,and the model is refined and the target model is mined.(2)Aiming at the mining of business processes,this paper proposes a method of mining hidden transition of business processes based on regions.Mining of hidden transition can better restore the business process model.There are some limitations in the method of mining in the accurate search of hidden transition.A method of mining hidden transition of business process based on region is presented in this paper.First,execution event logs are given.The interactive net(initial)process model of module net and feature net is constructed based on behavioral profile theory.Then,preprocessed fragment record event logs are transformed into a transition system to find non-trivial regions according to region theory,a fragment sub-model with hidden transition is built.The sub model is merged with the initial model,and the target model with hidden transition is mined.(3)For the mining of the infrequent behavior of business process,the method of conditional infrequent behavior mining and optimization under data awareness is proposed.Feasible traces of the reference model are inquired,and attribute values are added to infrequent traces.Data dependency values between the features that interact are calculated,and the infrequent behavior is classified.Moreover,the module net and the feature net are constructed according to the optimized log.The module net interacts with feature nets,and the business process optimization communication model of the feature net and the module net is mined.(4)For the mining and optimization of infrequent behavior in business process conditions,A method of mining conditional infrequent behaviors based on communication behavioral profiles was proposed.Conditional infrequent behavior refers to the behavior recorded by infrequent event traces with attribute values.Mining infrequent behavior from the event log is one of the main contents of business process optimization.Existing methods removed low frequency behavior,but also less consideration of conditional infrequent behavior based on data-flow perspective between different module nets.Based on this,the paper presents a method of mining conditional infrequent behavior based on communication behavioral profile.This method based on the communication behavior profile theory between module nets.Firstly,through a given business process source model,we can query its executable event log and find the infrequent event traces.Adding the relevant attributes and attribute values to the infrequent event traces to get the conditional infrequent traces.Secondly,calculating conditional dependent values of communication features between different the module nets.We can determine whether conditional infrequent traces are deleted or retained,the optimized event log is given.The business process optimization communication model is mined.