Making compliance measures actionable: a new compliance analysis approach

Abstract

Process mining can be used to measure the compliance between the actual behavior and the designed process. Traditionally, a single figure expressing the overall process compliance has only limited value to managers trying to improve their processes. This article proposes a new compliance methodology which first clusters the event log into homogeneous groups of event traces and then computes the compliance degree for each cluster separately. Additionally, each cluster is profiled by means of case information, which allows the discrimination between less and more compliant parts of the process. The benefits of this new compliance methodology in a business context are illustrated by means of a case study.