A Technique for Collaboration Discovery

In the last years, researchers have contributed to the process mining domain with several techniques and tools supporting the discovery of business processes. Almost all these contributions rely on event logs stored in the information systems of single organizations. In contrast, the discovery of collaborative scenarios where the information systems are distributed among different interacting organizations has been disregarded. In this context, we propose a novel technique for discovering collaboration models from sets of event logs stored in distributed information systems. Given the distributed logs of interacting organizations, the technique discovers each organization’s process through one of the available algorithms introduced by the process mining community. It also analyzes the logs to extract information on messages exchange. This information permits automatically combining the discovered processes into a collaboration diagram representing the distributed system’s behavior and providing analytics on messages exchange. The technique has been implemented in a tool and evaluated via several experiments.