Generating LTL Formulas for Process Mining by Example of Trace

Abstract

Process mining enables efficient and exhaustive analysis of business processes based on event data. Process mining tools such as LTL Checker allow users to verify temporal properties of business processes for traces by providing a description based on linear temporal logic (LTL). However, it is difficult for many users to understand and use LTL-based mathematical notation. Therefore, there is a need for a method to describe temporal properties even for those who are not familiar with mathematical notation. In this study, we proposed a method based on the satisfiability problem (SAT) for event logs in XES format used in process mining, and it is verified how well it can automatically generate temporal properties in business processes. We conducted experiments using two types of event logs to demonstrate the effectiveness of the proposed method.