A Two Phase Approach for Process Mining in Incomplete and Noisy Logs
The purpose of process mining is extracting knowledge from even logs recorded in executive information systems. In many real life logs, too many log instances are needed for the mining approach to work properly. In existence papers about process mining just complete or parallelism tasks with large logs were discussed but in this paper noisy and incomplete logs similarly tested. Therefore, another definitions, metrics and algorithms are required to mine event logs with not enough instances. In this paper, a probabilistic approach is proposed to mine event logs when the number of instances is limited In comparison with many existing approaches, based are the results of our experiments, the proposed approach is very robust in mining process logs with high degrees of parallelism, incompleteness and noise.
Keywords: process mining, incomplete, noisy, log, parallelism
Download Full-Text
ABOUT THE AUTHORS
Roya Zareh Farkhady
Department of Computer Science, Bostanabad Branch, Islamic Azad University, Bostanabad, Iran
Seyyed Hasan Aali
Department of Computer Science, Bostanabad Branch, Islamic Azad University, Bostanabad, Iran
Roya Zareh Farkhady
Department of Computer Science, Bostanabad Branch, Islamic Azad University, Bostanabad, Iran
Seyyed Hasan Aali
Department of Computer Science, Bostanabad Branch, Islamic Azad University, Bostanabad, Iran