Related Experiment Videos
Detection and mitigation of abusive web traffic using convolutional neural networks
Farkhanda Athar1, Akmal Shahbaz2, Mansoor Qadir3
1Department of Computer Science, The Islamia University of Bahawalpur, Bahawalpur, 63100, Pakistan.
Scientific Reports
|May 21, 2026
Summary
Online Abusive Traffic Finder (OATF) detects malicious Traffic Distribution Systems (TDSs) and associated threats. A CNN model achieved 91.92% accuracy in identifying abusive webpages, enhancing web security.
Area of Science:
- Cybersecurity
- Web Security
- Malware Analysis
Background:
- Traffic Distribution Systems (TDSs) are increasingly exploited by illicit websites to generate malicious user traffic.
- These systems facilitate various abusive activities, including phishing, scams, ad fraud, and social engineering attacks.
Purpose of the Study:
- To introduce the Online Abusive Traffic Finder (OATF), an advanced system for investigating and evaluating abusive TDSs and their associated threats.
- To enhance web security by understanding and mitigating evolving malicious traffic distribution strategies.
Main Methods:
- Collected 10,746 webpages from diverse sources like URL shorteners, typosquatting sites, and PhishTank over one month.
- Employed semi-automatic labeling and automated content examination, incorporating screenshots, browser, and content logs.
- Integrated a Convolutional Neural Network (CNN) classifier for automated abusive webpage detection using visual features.
Main Results:
- The CNN-based classifier achieved a high accuracy of 91.92% in identifying abusive webpages within the OATF framework.
- The study provided deeper insights into the operational behavior of abusive traffic ecosystems.
Conclusions:
- The OATF system offers an effective approach to detecting and combating abusive TDSs.
- The findings contribute to improving overall web security against sophisticated malicious distribution tactics.