“…They usually deal with typical problem of event data representation, including, for instance, noise (missing or duplicated records) and incompleteness, i.e., a finite event cannot cover all possible process executions. The paper [6] also stressed that an event log of a multi-agent system requires the additional inspection of agent behavior, since the direct discovery from a multiagent system event log produces process models the structure of which does not explicitly reflect agent behavior as sub-models and agent interactions as distinguished nodes. This happens because the concurrent execution of relatively independent agents leads to a wide range of possible traces recorded in an event log of a multi-agent system.…”