Determining the number of trace clusters: a stability-based approach

Given the complexity of real-life event logs, several trace clustering techniques have been proposed to partition an event log into subsets with a lower degree of variation. In general, these techniques assume that the number of clusters is known in advance. However, this will rarely be the case in practice. Therefore, this paper is the first to present an approach to determine the appropriate number of clusters in a trace clustering context. In order to fulfil this objective, a stability-based method for identifying the most appropriate number of trace clusters is proposed. The method involves the design of tailored resampling strategies and cluster similarity metrics. Regarding practical validation, our approach is tested on multiple real-life datasets to investigate the workings of the different components. Our results suggest that our method is successful in identifying the right number of trace clusters.status: publishe