An Experimental Evaluation of Passage-Based Process Discovery

Abstract

In the area of process mining, the ILP Miner is known for the fact that it always returns a Petri net that perfectly fits a given event log. Like for most process discovery algorithms, its complexity is linear in the size of the event log and exponential in the number of event classes (i.e., distinct activities). As a result, the potential gain by partitioning the event classes is much higher than the potential gain by partitioning the traces in the event log over multiple event logs. This paper proposes to use the so-called passages to split up the event classes over multiple event logs, and shows the results are for seven large real-life event logs and one artificial event log: The use of passages indeed alleviates the complexity, but much hinges on the size of the largest passage detected.