Streaming Conformance Checking with Log Skeleton in Python



Alessandro Berti



+49 241 80 21912



BSc Thesis Project

Title: Streaming Conformance Checking with Log Skeleton in Python

Student: Le Hieu

1st Supervisor: Prof. Dr. Wil van der Aalst (i9)

2nd Supervisor: Prof. Dr. Ulrik Schroeder


Alessandro Berti (M.Sc.) (i9)

Dr.Ing. Merih Seran Uysal (i9)


In recent years, process mining grew rapidly in both academia and industry. One of the most important types of process mining is conformance checking, which investigates if a process in real life is running as expected. However, the conventional conformance checking, i.e., analyzing the data in an a posteriori fashion does not allow to detect the deviations at the moment they occur, hence, it causes drawbacks in several cases. Our work presents a novel way of doing streaming conformance checking using the constraints derived from a log skeleton model. The log skeleton is a declarative process model which is considered as one of the most accurate approaches to process discovery/classification. The proposed method examines an event received from the event stream if the execution violates several constraints declared in the log skeleton model. To evaluate our work, based on an event log, we compare the results produced by the conventional conformance checking technique with the results generated by the streaming conformance checking algorithm. The evaluation shows the similarity and indicates the validity of our study.