Abstract
Automated process discovery is a branch of process mining that allows users to extract process models from event logs. Traditional automated process discovery techniques are designed to produce procedural process models as output (e.g., in the BPMN notation). However, when confronted to complex event logs, automatically discovered process models can become too complex to be practically usable. An alternative approach is to discover declarative process models, which represent the behavior of the process in terms of a set of business constraints. These approaches have been shown to produce simpler process models, especially in the context of processes with high levels of variability. However, the bulk of approaches for automated discovery of declarative process models are focused on the control-flow perspective of business processes and do not cover other perspectives, e.g., the data, time, and resource perspectives. In this paper, we present an approach for the automated discovery of multi-perspective declarative process models able to discover conditions involving arbitrary (categorical or numeric) data attributes, which relate the occurrence of pairs of events in the log. To discover such correlated conditions, we use clustering techniques in conjunction with interpretable classifiers. The approach has been implemented as a proof-of-concept prototype and tested on both synthetic and real-life logs.