On the Application of Locality Sensitive Hashing to Behavioral Web Analytics

In today’s constantly connected world, the dependence on web-based technologies is ubiquitous, creating opportunities for both malicious and benign activity. As a result, it is essential that we be able to identify users on the web. Although simple methods, such as tracking a user by userid or by IP address exist, these methods can easily be evaded if the user so desires, by creating multiple ids or operating from different IP addresses. However, due to habit, a user does not often change the way he or she browses the web, such as the time of day that she visits various genres of pages. In this project, we evaluate locality sensitive hashing (LSH) to uniquely identify and authenticate users based only on the day of the week, time of day, and genres of websites. In addition, we provide a novel extension of LSH, Mode Closest Hash (MCH).