Analysis of Event Logs: Behavioral Graphs

摘要

Analysis of event logs is very important discipline used for the evaluation of performance and control-flow issues within the systems. This type of analysis is typically used in process mining sphere, where information systems, for example workflow management systems, enterprise resource planning systems, customer relationship management, supply chain management systems, and business to business systems record transactions and executed activities in a systematic way. Social network analysis takes part of process mining techniques, focused on activity performers, on users. The authors present a new approach to analysis of user behavior in the systems. The approach allows to find behavioral patterns and to find groups of users with similar behavior. An observer can obtain relations between the users on the basis of their similar behavior. The visualization of relations between the users is then presented by so called behavioral graphs. The approach was tested for event log analysis of a virtual company model developed as a multi-agent system by modeling environment MAREA.