A Generator of Synthetic Access Logs that Contain Realistic User Behavior Patterns

Generating high quality synthetic data for testing algorithms and system implementations is challenging. This research designed and developed a tool called QUAlity Synthetic Information Log Generator (Quasi-Log) to facilitate the development and testing of a new series of Information Discovery Systems (IDS) that focus on detecting User Behavior Patterns to improve the quality, security and potential profitability of systems. Quasi-Log is a powerful and highly configurable synthetic event log generation engine with advanced support for semantic attributes and value sets, attribute correlations, event sequence pattern definitions and user behaviour pattern definitions. In this paper we discuss the design of Quasi-Log and provide empirical results to demonstrate that our tool is effective at generating semi-random log files that exhibit these interesting properties.