Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{182123,
        author = {Chejarla Vinod Kumar and Dr. Shaik Shakeer Basha and Dr. Mohammed Abbas Qureshi},
        title = {ObscuraNet: Ensemble-Learning-Based Dissection of Onion Service Traffic within Anonymity Networks},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {2},
        pages = {923-927},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=182123},
        abstract = {The increasing adoption of privacy-preserving technologies has transformed the landscape of online anonymity, particularly through the Tor network and its support for Onion Services. While these services aim to ensure user privacy, their anonymizing capabilities are often exploited, prompting governmental and institutional interest in traffic analysis for de-anonymization. In this study, we explore the classifiably of Onion Service traffic within the broader spectrum of Tor traffic. Our work presents three core contributions. First, we develop a high-precision classification pipeline capable of isolating Onion Service traffic from general Tor traffic, achieving over 99% accuracy using advanced feature-based techniques. Second, we rigorously assess the robustness of our classifier against traffic obfuscation strategies, simulating realistic evasion tactics. Our findings indicate a notable drop in classification performance—up to 15%—when such obfuscations are applied, underscoring the dynamic challenge of traffic concealment. Third, we perform an in-depth analysis of feature importance, revealing the most influential attribute combinations driving classifier decisions. As an extension, we integrate ensemble learning approaches—namely AdaBoost and Voting Classifiers—into our framework, which collectively enhance classification performance and achieve up to 100% accuracy under standard conditions. This highlights the promise and limitations of machine learning in discerning anonymized traffic and has profound implications for both traffic monitoring and privacy preservation.},
        keywords = {Tor network, Onion Services, Traffic classification, Ensemble learning, AdaBoost, Voting Classifier, Website Fingerprinting, Traffic obfuscation, Anonymity networks, Random Forest, Support Vector Machine (SVM), K-Nearest Neighbors (KNN).},
        month = {July},
        }