Pattern Mining in System Logs: Opportunities for Process Improvement

Enterprise systems implementations are often accompanied by changes in the business processes of the organizations in which they take place. However, not all the changes are desirable. In “vanilla” implementations it is possible that the newly operational business process requires many additional steps as “workarounds” of the system limitations, and is hence performed in an inefficient manner. Such inefficiencies are reflected in the event log of the system as recurring patterns of log entries. Once identified, they can be resolved over time by modifications to the enterprise system. Addressing this situation, the paper proposes an approach for identifying inefficient workarounds by mining the related patterns in an event log. The paper characterizes such patterns, proposes a mining algorithm, and rules for prioritizing the required process improvements.

• Type: Conference
• Year: 2009
• Where: CAISE
• Authors: Dolev Mezebovsky, Pnina Soffer, Ilan Shimshoni
• Keywords: CAISE 2009, Enterprise System, Enterprise Systems Implementations, Event Log, Information System