A Study on Browser Fingerprinting Uniqueness Using Clustering Methods and Entropy Validation

Browser fingerprint is often linked to privacy as it is a method to gather data about the browser's configuration to identify the user. The browser’s configurations, which are also known as attributes, are the keys to make the user to be identified. Web browsers explicitly disclose information about the host system to websites by making it available to them, such as attributes like the screen resolution, local time, or operating system (OS) version. Since each of the browsers has different attributes that make each unique, it is essential to understand the attributes well. This research paper emphasizes the method of collecting data for browser fingerprinting and ensuring the acquisition of fingerprint data without compromising personal information. One of the research motivations is to transform this data into an easily accessible raw dataset for the industry's utilization in future research projects. Additionally, the study explores the potential use of Shannon Entropy to unveil distinctive attributes in browser fingerprinting, revealing that higher entropy values correlate with more distinct and recognizable fingerprints. The other purpose is to discover which attribute produces the highest unique value using the clustering algorithm. Experiment results showed that if the attribute is unique, it will be hard to cluster into groups. This can be proved by using a clustering algorithm where the unique attributes will have a high value in the incorrectly clustered instances because it is harder to be clustered.