On Business Process Redesign and Configuration: Leveraging Data Mining Classification & Outliers and Artifact-Centric Process Modeling

In today's e-business, data increasingly capture many behavioral aspects of our daily life, e.g. operational patterns in production lines, human shopping habit, medical treatment, supply-chain management. We formally define the behavior of tasks (of the business process in question) in connection with conceptually described business artifacts that are manipulated by the tasks. We relate this representation to data mining and nonfunctional requirements to better define the context for process redesign (i.e, design-time changes) and process configuration (i.e, run-time changes). Furthermore, we propose algorithms for business-process redesign and configuration based on the above context. Our research is positioned in the field of business process flexibility – a term coined by van der Aalst [1] in his thorough literature survey on business processes. We validate our framework using real-life data provided by our industrial partners.