Coupled Hidden Markov Model for Process Discovery of Non-Free Choice and Invisible Prime Tasks

With increasing numbers and complexity of processes stored in the event log on a system, a solution to analyze these processes more easily is a formation of the process model. There are algorithms of process discovery that offered ways to formed process models based on data in the event logs. During the forming of process models, several problems arise, such as detecting non-free choice and invisible prime tasks. This research offers a method that utilizes Coupled Hidden Markov Model to modeling processes with non-free choice and invisible prime tasks in incomplete event log. This method was compared with Alpha$ algorithm to evaluate the quality of the method. The evaluation proves that the method of this research can depict process model containing non-free choice and invisible prime tasks. This output of the evaluation also indicates that a process model produced by the method with dealy Coupled Hidden Markov Model has higher fitness and precision than the model produced by Alpha$ algorithm.