Workflow Pattern mining and Its applications

• Main Authors: Guan-Lin Lai, 賴冠霖
• Other Authors: Ou-Yang Chao
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/42905275672377619217

碩士 === 國立臺灣科技大學 === 工業管理系 === 98 === Process oriented ERP system will store the activities records in the event logs. In general, there exist certain interrelationships among these activities called “business activities rules”. However, it would be difficult for manager to identify these embedded business rules from the event logs. Therefore, if a mining approach can be developed to discover these rules, it can be used as reference information in the management level function such as auditing. In this research, an approach based on genetic process mining algorithm to discover the business rules will be developed. The logical relationships among the process activities will be identified based on the causal matrix generated from genetic mining algorithm. The identified relationships will then be transformed to Petri-net based business activities rules. A three-level developing framework will be proposed. In the conceptual level, the relationships among the process activities, event logs, genetic mining and business activities rules will be discussed, as well as the types of logical relationships among the activities will be studied. Then, in the design specification level, an algorithm will be proposed to discover the single-level and multi-level business rules from the causal matrix. In the implementation level, an event log for activity changing process from SAP R/3 will be used to as an empirical test. It is expected that certain business rules about the changing activities can be discovered.