Defining a Method to Evaluate the Semantic Quality of Process Models on the Fly
Master’s thesis, University of Innsbruck, Department of Computer Science, 2015.
The wide-spread application of business process management techniques raised the importance
of business process modeling. Process models are providing a common understanding
of the utilized business processes in an organization. Still, industrial process
models in large process repositories are affected by quality issues. The scientific community
invested comprehensive effort in research towards process model quality focusing
on final process models. A new stream of research proposes an assessment of quality
throughout process model creation also denoted as the process of process modeling. Still,
methods for assessing the process model quality with respect to the process of process
modeling are only available to a limited extent so far.
This thesis strives on addressing this gap by establishing an approach for assessing
the quality throughout the process of process modeling in the light of semantic correctness.
For that, existing approaches proposed in literature are screened and a promising
approach is chosen. In particular, the widely used method of behavioural profiles is
selected to serve as foundation. Then, this approach is adapted to deal with the peculiarities
of incomplete process models gathered during process model creation. Finally,
the extended functionality is applied to process model created during experiment sessions
to evaluate the effect of the suggested implementation. In the process, two process
modeling experts assessed the process models gathered during experiment sessions with
their own approaches and provided these findings to facilitate a validation of the method.
In conclusion, this thesis provides a promising approach for assessing semantic quality
of incomplete process models on the
y by adapting an existing method to the requirements
of process models gathered during process model creation at a semantic level. In
particular, the approach facilitates a quick and precise detection of modeling errors in
regard to the moment of emergence.
More publications by