Application of Tree-structured Data Mining for Analysis of Process Logs in XML format

Process logs are increasingly being represented using XML based templates such as MXML and XES. Popular XML data mining techniques have had limited application to directly mine such data. The majority of work in the process mining field focuses on process discovery and conformance checking tasks often utilizing visualization and simulation based techniques. In this paper, an approach is proposed within which a wider range of data mining methods can be directly applied on tree-structured process log data. Clustering, classification and frequent pattern mining are used as a case in point and experiments are performed on publicly available real-world and synthetic data. The results indicate the great potential of the proposed approach in adding to the available set of methods for process log analysis. It presents an alternative where process model discovery is not the pre-requisite and a variety of methods can be directly applied.