Algorithm Of Incomplete Log's Processing Mining

Modern enterprises continuously improve their business process operation and introduce Process-Aware information systems,such as workflow management system and ERP to meet the needs of high productivity and superior service.Process mining can utilize the information system logs saved by the system implementation process to excavate the objective knowledge.These discovered objective knowledges contribute to the deployment of new business systems,or as a feedback tool to improve existing business processes.This paper mainly studies the control flow and organizational perspective in process mining.Firstly through studying domestic and foreign research situation of process mining field,have a deep understanding of different algorithms of process mining and find that there are incomplete log problems in the perspective of mining control flow.In order to overcome this problem,this article defines the prior knowledge of business domain personnel as a set of constraint rules,based on the dependency graph,presents a process mining algorithm for incomplete logs and some key problems are discussed in detail,such as the structure of the initial dependency graph,meeting negative and positive constraints,binding operation.Then,in the perspective of the mining organization,the traditional process mining algorithm has defects in the ability to find the multiple role identity resources.On the basis of a social network,we can get the partition of resource community by optimizing modularity about network.Meanwhile we can also transform social network to weighted community network,find out the neighbor with largest fitness of community and then form resource role network,the role that resource play in different communities can be found.Finally,the correctness of the algorithm is verified by experiments.In the case of incomplete log,this paper classifies the relationship between negative path constraints and reduces the number of edges that should be removed or added.The effective of process is improved in the operation of satisfying positive constraint by weight-best updating method.Moreover,the binding operation reduces the generation of the non-original log track,and finally the process model is more appropriate to the original model.In terms of mining resources in logs,the algorithm in this paper solves the problem that traditional algorithms can not exploit overlapping communities,after the partition of communities,the modularity of network is higher,and it can find resources with multiple roles.