Analyzing compliance of service-based business processes for root-cause analysis and prediction

Automatically monitoring and enforcing compliance of service-based business processes with laws, regulations, standards, contracts, or policies is a hot issue in both industry and research. Little attention has however been paid to the problem of understanding non-compliance and improving business practices to prevent non-compliance in the future, a task that typically still requires human interpretation and intervention. Building upon work on automated detection of non-compliant situations, in this paper we propose a technique for the root-cause analysis of encountered problems and for the prediction of likely compliance states of running processes that leverages (i) on event-based service infrastructures, in order to collect execution evidence, and (ii) on the concept of key compliance indicator, in order to focus the analysis on the right data. We validate our ideas and algorithms on real data from an internal process of a hospital.