Event Correlation Analytics: Scaling Process Mining Using Mapreduce-Aware Event Correlation Discovery Techniques

Abstract

This paper introduces a scalable process event analysis approach, including parallel algorithms, to support efficient event correlation for big process data. It proposes a two-stages approach for finding potential event relationships, and their verification over big event datasets using MapReduce framework. We report on the experimental results, which show the scalability of the proposed methods, and also on the comparative analysis of the approach with traditional non-parallel approaches in terms of time and cost complexity.