Content area
Full Text
Abstract
The Business Process Management System (BPMS) has received more attention as companies increasingly realize the importance of business processes. However, traditional BPMS focuses mainly on correct modeling and exact automation of the process flow and pays little attention to the achievement of the final goals of improving process efficiency and process innovation. During and after execution of processes, BPMS usually generates much process log data in which numerous meaningful rules and patterns are hidden. In the present study we employ a data mining technique to extract useful knowledge from the complex process log data. A data model and a system framework for process mining are provided to help understand the existing BPMS. Experiments with the simulated data demonstrate the effectiveness of the model and of the framework.
Keywords
business process, business process mining, process pattern, decision tree
1. Introduction
Companies have recently been faced with internal and external challenges at the same time they try to cope with a keenly competitive business environment. Their efforts to produce and deliver numerous types of products and services have not only increased the number of their processes, but also complicated their processes as well. Moreover, shortened product life cycles require frequent process changes. Since the early 1990s, the advent of new business paradigms, such as business process reengineering (BPR) and continuous process improvement (CPI) have altered fundamental thinking about organizational structure. These concepts remind companies that their information systems must provide integrated coordination and control of processes as well as deal with individual tasks that facilitate the introduction of business process management (BPM). The BPM system is an information system that automates the execution and management of business [1] [2]. The BPM system can automatically control the sequence of work, expedite a processing task by electronic documents, reduce unnecessary paperwork, and monitor the status of a process. Unfortunately, however, most current commercial BPM systems fail to capture the full potential of the technology available because they are mainly designed only to automate the execution of business processes. Although a few studies have tried to use business process data for simple analysis of process log data [3] [4], little effort has been made to use a systematic framework to analyze and evaluate business processes. In this paper...