Supporting Provenance and Data Awareness in Exploratory Process Mining

Abstract

Like other analytic fields, process mining is complex and knowledge-intensive and, thus, requires the substantial involvement of human analysts. The analysis process unfolds into many steps, producing multiple results and artifacts that analysts need to validate, reproduce and potentially reuse. We propose a system supporting the validation, reproducibility, and reuse of analysis results via analytic provenance and data awareness. This aims at increasing the transparency and rigor of exploratory process mining analysis as a basis for its stepwise maturation. We outline the purpose of the system, describe the problems it addresses, derive requirements and propose a design satisfying these requirements. We then demonstrate the feasibility of the central aspects of the design.