Predicting System Behavior through Deep Learning of Process Mining Models and Generalization Metric

With the ongoing digitization of industries and the rising number of interconnected devices, an increasing amount of system recordings are created. Process mining describes a set of techniques that help with the automatic discovery of mathematical models from such system recordings. The generated mathematical models represent the dynamic behavior of the system. The analysis of the discovered models might reveal important knowledge about the system behavior that is otherwise difficult if not impossible to obtain. Process mining techniques mainly rely on discrete event system theories in constructing the corresponding mathematical models. Therefore, they perform poorly when continuous measures such as time and probabilities of the system events need to be considered. On the other hand, such measures can be effectively modeled and predicted through deep learning. This dissertation focuses on the symbiosis of process mining and deep learning to confidently model the underlying system and predict its behavior by incorporating the advantages of both disciplines. The potentials of the proposed methodologies are demonstrated with an application in the healthcare domain.