Detecting Model Changes in Organisational Processes: A Cloud-Based Approach

Process mining techniques extract knowledge from event logs within organizations to understand and improve the behavior of their business processes. These techniques utilize a wide range of methods to automatically generate process models from event log data, simplify these models, calculate various indicators to optimize performance, and visualize and explain model behavior. However, these techniques often treat process models as static entities, despite the inherent dynamic nature of processes. Commercial platforms frequently lack the ability to detect and describe changes (also known as concept drift) in the models, which can impact the conclusions and results derived from process mining. This paper presents the INSIDE-TUTTO project, which has developed a concept drift detection algorithm for application in business organizations and transition to the commercial market through Inverbis Analytics. The original algorithm was not designed to operate in real-world scenarios with large volumes of data. By combining distributed architectures and the cloud computing paradigm, the algorithm was evolved into a commercial version deployed within Inverbis Analytics’ Azure-based technological infrastructure.

keywords: