Algorithm Research Of Workflow Mining In Noise Environment

Today, workflow mining is one of the research areas in workflow. Workflow mining aimsat extracting information from event logs to capture the business process as the enterpriseinformation system is being executed.In the process of designation about the traditional information system, such asmanagement information system (MIS), enterprise resource planning (ERP), customerrelationship management (CRM), the designers would be affected by their own subjectivefactors. They design the workflow models which are not consistent with the need of actualwork. For this reason, workflow mining came into being. The designers use the result ofworkflow mining as a reference for the business process redesign. Then they design aworkflow model which is consistent with the need of actual work.Many researchers have proposed solutions for workflow mining in the research area ofworkflow mining. However, many of the solutions have defect about workflow mining. The-algorithm is the representative algorithm about workflow mining. The paper will analyzethe defect of it and discuss the noise about the workflow mining. Then itillustrates the D/Ftable (Dependency/Frequency table), the noise factor, the threshold factor and the causalityfall factor, and use them to mine the causal relationship under the noise environment. Afterthat, it introduces methods which solve the problem of concurrent relationship and looprelationship by the threshold factor. The paper proposes a new algorithm to solve thenon-free-choice structure, which is the one of the defect of-algorithm. The new algorithmdecides whether a non-free-choice structure exist or not by analyze the OR-join structure inthe workflow event logs.The paper introduces the+-algorithm, which is the improved-algorithm, and make itbecome a plug-in of the ProM.