Interactively exploring logs and mining models with clustering, filtering, and relabeling

Real-life event logs often contain many data quality issues, which obstruct existing discovery algorithms from discovering meaningful process models and process analysts from conducting further process analysis. In this paper, we present an integrated tool that provides support for dealing with three of these data issues: logs comprising recordings of multiple heterogeneous variants of a process; traces containing multitude of deviating events in an infrequent context; event labels being imprecise. The tool is called Log to Model Explorer and helps users in interactively and iteratively exploring and preprocessing a log by clustering, filtering and event relabeling, enabling them to discover more meaningful process models.