Enhancing change mining from a collection of event logs: Merging and Filtering approaches

Abstract

Abstract An event log is the key element of all change mining and process mining approaches. Those approaches bridge the gap between conventional business process management and data analysis techniques such as machine learning and data mining. In this day, companies and business organizations usually use a family of business processes that may face different variations and adjustments. Still, those processes are widely identical, with a slight difference in specific points. Consequently, performing a process mining or a change mining for each process will be a redundant task. The use of a configurable process model is a practical solution for redundancy problem. Thus, the process mining areas such as discovering verifying the conformity of a business process and enhancing processes, are reduced considerably. However, the configurable process models and the variability concept are rarely introduced in change mining approaches. The existing methods that analyse and manage event logs do not then consider the variability issue. Therefore, the fact of using a collection of event log becomes a challenging task. Our proposed approach is to merge and filter a collection of event logs from the same family with respect to variability. Our goal is to enhance change mining from a collection of event logs and detect changes in variable fragments of the obtained event log.