Bot-like Behavior Detection in Online Banking
Yuan Wang, Liming Wang · 2019
To ensure the security of online banking, banks often analyze user behavior and dig out abnormal behaviors. User behavior analysis relies heavily on user behavior models established from user activity data. However, if researchers are not aware of the behaviors of bots in the data, then the built user behavior model may be vulnerable, and even unable to effectively detect abnormal behavior. In this work, we analyze the user behavior from the perspective of the inter-request time interval and find that bot-like behaviors exist in the online banking system. Then, different classification algorithms entropy-based are used to detect bot-like behavior in online banking transactions. The results show that Renyi entropy is superior to Shannon entropy in distinguishing bot behavior and human behavior in real data.