Modeling parallel business process using modified time-based alpha miner

One part of process mining is process discovery. A process discovery collects the information from an event log and obtains the business process model representing the behavior of activities based on the information. A business process model contains activities and their relations, specifically sequential and parallel relations. In this paper, Modified Time-based Alpha Miner algorithm, which is a modification of the Alpha Miner algorithm is presented. The proposed algorithm considers both the sequence of activities and time interval information from the event log to discover parallel business process model. Time interval information from event log can be used to determine parallel gateway AND and OR of business process. The experimental results show that the process models formed by parallel gateway AND and OR are able to be discovered using the proposed algorithm, whereas using the same process models, the existing Alpha algorithm determines them as a parallel gateway AND. The results also show that the proposed algorithm needs less traces than the existing Alpha Miner to discover process models. To evaluate the proposed algorithm and Alpha Miner, the comparison of fitness value is presented. It also clearly states that our proposed algorithm can present better results rather than the existing Alpha Miner algorithm.