Web Bot Detection Using Keyboard Behavioural Analysis

Ever increasing technology and internet availability resulted in a huge rise in web bots. Users utilize web bots with good or malicious intent. The increase in web bot traffic has raised concerns for the safety and security of the web. To address this issue, there was a rise in the development and implementation of bot detection mechanisms. To ensure the safety and security of the web, various methodologies have been applied, including algorithms based on temporal and behavioral characteristics, CAPTCHAs, etc. With advancements in bots mimicking human fingerprints, mouse movements, and feeding CAPTCHAs, it became very easy to evade these detections. In this paper, an additional layer of security is introduced that utilizes keyboard behavior analysis to detect bots. The proposed algorithm works on collecting and storing data related to each keystroke, which includes records based on timestamps, key names, key hold time, and key time difference. The algorithm processes the recorded data through various conditions and parameters to conclude the detection. This algorithm works on top of the other detection mechanisms, like weblog and mouse movement detection. The proposed algorithm is implemented on a publicly provided data set to measure its effectiveness and accuracy. The findings prove that the algorithm works as an effective layer for detecting bots through the input mechanism.