Model business process improvement by statistical analysis of the users' conduct in the process

Abstract

Process Mining is a research area that meets the gap between business processes and various IT systems. Most of the works in this area focus on the control flow perspective, while very few of them address the organizational aspect. The organization perspective of process mining supports the discovery of social network within organization by analyzing events logs recorded during real process execution. For process owners, it is very important to know how users perform their activities in process. In this paper, we introduce a process discovery method that combines an organizational perspective with probabilistic approach. Combining these two approaches we are able to fit distribution of users work and distribution of instance generation in process. We use different statistic methods like Cullen and Frey graph, Kolmogorov-Smirnov statistic test, Carmen-von-Mises statistic test and Anderson-Darling statistic test. After finding appropriate distribution we estimate its parameters. Research conducted and presented in this paper reveals that the information about users behaviour in process is significantly useful in further analysis: in simulations, to identify bottlenecks, to improve productivity of resource management and to identify task complexity in process.