Inducing declarative logic-based models from labeled traces

Abstract

In this work we propose an approach for the automatic discovery of logic-based models starting from a set of process execution traces. The approach is based on a modified Inductive Logic Programming algorithm, capable of learning a set of declarative rules. The advantage of using a declarative description is twofold. First, the process is represented in an intuitive and easily readable way; second, a family of proof procedures associated to the chosen language can be used to support the monitoring and management of processes (conformance testing, properties verification and interoperability checking, in particular). The approach consists in first learning integrity constraints expressed as logical formulas and then translating them into a declarative graphical language named DecSerFlow. We demonstrate the viability of the approach by applying it to a real dataset from a health case process and to an artificial dataset from an e-commerce protocol. © Springer-Verlag Berlin Heidelberg 2007.