Study On The Methodology Of Business Process Analysis Based On Process Mining

The consolidation and promotion of the core competitiveness of enterprises can not be separated from the business process management.The business process data recorded in the information systems provides a new opportunity for business process management,while the traditional business process management methods can not make use of the existing business process data effectively.In recent years,the growing process mining technology aims to extract useful knowledge for discovering,monitoring and improving business processes from business process data.Although the theoretical research of process mining is becoming more and more mature,there are still some defects in the aspects of process mining framework,case information extraction,process modeling and diagnosis results analysis.Considering these problems,on the basis of literature collection of process mining case studies,a three-dimensional framework model of business process analysis based on process mining is proposed in this study,which includes event log dimension,process modeling dimension and multi-perspective dimension.By comparing with the existing process mining framework models,the effectiveness of the proposed framework model in this dissertation is verified.The specific findings and contributions for the three dimensions are as follows:Firstly,in the dimension of event log,based on text classification techniques,a case information extraction method for natural process texts is proposed.Current case information extraction methods focus on extracting named entities and relationships contained in process texts.When named entities in process texts cannot represent or be abstracted as target activity types,existing case information extraction methods will fail.This problem is effectively solved by introducing text categorization algorithms into the process of case information extraction.Based on some complaint handling service process data of D Enterprise in 2017,this dissertation conducts a case study to verify the effectiveness of the proposed method.Secondly,in the dimension of process modeling,a method of creating process model based on aggregating business activities by clustering algorithms and calculating the optimal cluster number for fuzzy mining algorithm is proposed.It effectively solves the problem that fuzzy mining algorithm aggregates mainly based on the importance of the nodes and the problem that the optimal cluster number is not easy to select.The proposed approach improves the business readability of the result model.Based on the emergency rescue data of coal mines from the second half of2006 to 2014,a case study is carried out to verify the effectiveness of the proposed method.Thirdly,in the dimension of multi-perspective,the Acci Map modeling analysis method is introduced into the analysis of multi-perspective diagnosis results of process mining and a method of business process analysis by creating Acci Map business process management model is proposed.This method integrates the multi-perspective diagnosis results based on process mining and can analyze the problems in the process of business process management from different system levels.It solves the dilemma that the analysis of mining results from different process mining perspectives focuses on the operation process level but not easy to analyze across system levels.Based on complaint handling service process data of D Enterprise from2012 to 2016,this dissertation conducts a case study to verify the effectiveness of the proposed method.This study extends the methodology of business process management based on process mining.Meanwhile,it provides specific direction and idea for the practice of business process management based on process mining,and explores the feasibility of process mining practice in enterprises.