Federated Process Mining: Exploiting Event Data Across Organizational Boundaries

Abstract

Many organizations use process mining to improve their internal processes. However, many processes span multiple organizations, e.g., organizations that form a supply chain or a global production network. In order to improve processes across different parties, one cannot assume a single overall event log. Organizations may not be willing to share event data and may use different identifiers and logging conventions. Federated process mining aims to tackle these problems by creating views on the cross-organizational processes such that analysis is possible. A federated event log transparently maps multiple organization-specific event logs into a single unified representation. Using such a federated event log, it is possible to apply a range of existing process mining techniques and tools. If organizations are not willing to share raw event data, they need to resort to event log abstractions. In such settings, federated process mining approaches merge these abstractions to create an overall view on the process. This paper provides a framework to reason about the different forms of federated process mining enabling inter-organizational process transparency and improvement.