Reconstructing meaningful workflow from transaction logs

Enterprises today rely on workflow more and more to accomplish business goals. A well designed process can contribute to better business operation and working efficiency. In practice, however, most organizations have no efficient workflow, either there is not at all from the very beginning, or lack domain experts who equip with the comprehensive landscape about the business. In order to conduct business more effectively, it is necessary to reconstruct workflows from transaction logs, and further, optimize the business via meaningful representations of these workflows. In this paper, we propose a method and corresponding toolbox to mine the workflows and vital patterns. Domain experts can use them in an interactive way to reconstruct meaningful business processes. We got very promising experimental results on two real applications datasets.