Exploring business process Deviance with Declare

This thesis introduces business process deviance mining, which belongs to the group of process mining, and gives an overview on multiple deviance mining approaches. After that we focus on deviance mining using discriminative patterns, which belongs to the group of sequential patterns mining techniques. In this work we propose new discriminative pattern mining algorithm based on the Declare language. We implemented the approach as a plug-in of the Process Mining tool ProM. We describe the whole proposed approach from the labelling of the event logs until building the classifier and classifying the test logs. In the end of the thesis we evaluate the effectiveness of our proposed algorithm on variety of experiments on event logs.