Synthetic Event Logs for the Benchmarking of Trace Clustering Methods

This data set contains generated event logs with case attributes for the comparison of trace clustering methods in process mining. Each event log consists of four different process behaviors with respect to the control-flow (i.e. the order of activities) and the data perspective (i.e. the case attributes). The set contains six different process models of varying complexity which were used to generate the event logs. Event logs were generated with different sizes, different noise levels (e.g. skipped activities, switching of activities, executing activities multiple times) and different number of attributes. Attributes were generated by sampling from a set of values and assigning these values to the different behaviors only. Additionally, there is some overlapping of attribute values.