A Algorithm for Detecting Concept Drift Based on Context in Process Mining

In the field of process mining, we require certain techniques that can detect process changes in dynamic systems automatically. Through detecting changes in the process we can adjust and optimize the overall process timely. These technologies are named concept drift detection of process mining. The traditional concept drift algorithm in process mining domain mostly have high time complexity, high space complexity and low recognition rate of concept drift and without using the context in process. We arises a new algorithm that based on context has high recognition rate of concept drift, low time complexity and low space complexity. The algorithm reduces the extraction and calculation of sample properties by using the stability of the processes before and after the changes. The using of the time parameter and staff parameter could enhance the sensitivity of process´s change.