Fateme Jafarinejad - AI & Distributed Systems Laboratory, School of Computer Engineering, Shahrood university of Technology Shahrood, Iran
Ali A. Pouyan - AI & Distributed Systems Laboratory, School of Computer Engineering, Shahrood university of Technology Shahrood, Iran
Morteza Zahedi - Web Mining & Pattern Recognition Laboratory, School of Computer Engineering,Shahrood university of Technology Shahrood, Iran

Extremely rapid growth of event data in many industries –including internet, social networks, cloud computing, are experienced. Process mining techniques aim to extract information from these event logs. However, overwhelming amounts of event data also provide new challenges that often existing process mining techniques cannot deal with. One of the important challenges in process mining is to discover a process model describing observed behavior in the best possible manner. This paper will focus on process discovery in the large . We suggest using a formal composition framework for process discovery. Trough this composition technique, two or more fine-grain process models can be integrated into a value-added coarse-grain process model. This composition reduces processing time of discovery procedure. We prove this algorithm to be able to discover a model of the whole system through a case study

Process Mining, Process Discovery, Petri Net Composition