A Machine Learning-Based Deanonymization Tool for Analyzing Anonymized Network Traffic

A Machine Learning-Based Deanonymization Tool for Analyzing Anonymized Network Traffic

Amit Kumar Sachan*1, Shikhar Kanaujia*2, Yashwant Kumar Sharma*3

Khushi Srivastava*4, Shivaditya Singh*5

*1Assistant Professor, Computer Science and Engineering, Babu Banarasi Das Institute of Technology and Management, Lucknow, Uttar Pradesh, India

*2,3,4,5 Student, Computer Science and Engineering (AI&ML), Babu Banarasi Das Institute of Technology and Management, Lucknow, Uttar Pradesh, India

Email – shikhar.kanaujia786@gmail.com

Abstract
Deanonymization tools are designed to unmask hidden identities or reconstruct obscured information, often leveraging advanced data analysis techniques. These tools have applications across cybersecurity, digital forensics, and law enforcement. The goal of this study is to develop a deanonymization tool that leverages metadata analysis, pattern recognition, and machine learning algorithms to identify and correlate digital footprints left across diverse platforms. This tool targets scenarios involving anonymized communication, obfuscated data logs, and masked web interactions, aiming to reveal underlying identities without compromising ethical boundaries. This project explores the concept of

deanonymization, a technique used to reveal the true identity of individuals or entities previously anonymized in data systems. In a college context, this topic is relevant due to the growing use of online systems for academic assessments, student interactions, and personal data storage. The project investigates the methods, tools, and challenges associated with deanonymization, particularly in relation to how data, once anonymized, can potentially be traced back to its source through advanced algorithms, metadata analysis, or pattern recognition. Our proposed tool employs multi-layered analysis, starting with metadata extraction from network packets, social media footprints, or anonymized datasets. By

correlating temporal, spatial, and contextual parameters, the system constructs potential user profiles. Next, machine learning models analyze behavioral patterns, leveraging clustering and classification algorithms to match anonymized entities with known datasets. The system also integrates Natural Language Processing (NLP) techniques for deanonymizing text-based communications, extracting linguistic traits that can link anonymous entities to real-world users. Ethical considerations and regulatory compliance are central to this project, ensuring the tool is used responsibly. Stringent safeguards are incorporated to prevent misuse, and its deployment is intended for authorized personnel in scenarios such as criminal investigations, fraud detection, and cyber threat mitigation. This deanonymization tool represents a significant step in advancing cybersecurity capabilities while emphasizing the balance between security needs and ethical considerations in the digital age.