Extracting configuration guidance models from business process repositories

Abstract

Configurable process models are gaining a great importance for the design and development of reusable business processes. As these processes tend to be very complex, their configuration becomes a difficult task. Therefore, many approaches propose to build decision support systems to assist users selecting desirable configuration choices. Nevertheless, these systems are to a large extent manually created by domain experts, which is a time-consuming and tedious task. In addition, relying solely on the expert knowledge is not only error-prone, but also challengeable. In this paper, we propose to learn from past experience in process configuration in order to automatically extract a configuration guidance model. Instead of starting from scratch, a configuration guidance model assists analysts creating business-driven support systems.