International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 11 | Nov 2024
p-ISSN: 2395-0072
www.irjet.net
Malware Link Detector: An Intelligent System for Real-Time Detection and Safe Redirection from Malicious URLs Rosini S1, Soniya S2, Trisha S3 ,Uma S4 1student,2student,3student, 4Assistant professor
Department of Computer Science and Engineering, Paavai Engineering College, Tamil Nadu, India ---------------------------------------------------------------------***--------------------------------------------------------------------Abstract-Malicious URLs are commonly used by cybercriminals to conduct phishing attacks, deliver malware, and exploit user
trust, posing significant risks to internet users. "Malware Link Detector" is a novel, machine learning-based system designed to detect and block such malicious URLs in real-time. In cases where a URL is identified as malicious, the system further supports users by suggesting alternative, safe links aligned with the original search intent. If no malicious activity is detected, the URL is safely opened in a new window. This dual approach not only protects users but also fosters safer browsing habits by offering educational redirection. This paper details the design, methodologies, implementation, and effectiveness of the Malware Link Detector, which shows high accuracy and efficiency in detecting malicious URLs while enhancing the browsing experience for users. Key Words: : Machine learning ,Malicious ,safer browsings,Dual Approach ,Educational Redirection, High Accuracy, Efficiency.
1.INTRODUCTION The internet has become integral to personal and professional life, with users relying on online platforms for various activities, from banking to social networking. As online interactions increase, so does the threat from malicious actors who exploit user trust through phishing links, malicious redirects, and malware-laden URLs. These URLs are frequently embedded in emails, social media posts, and advertisements, often disguised as legitimate links that lure users into clicking on them. Traditional cybersecurity defenses, such as blacklists, antivirus programs, and firewalls, provide a certain level of protection but often lack the ability to handle new and unlisted threats, known as zero-day attacks. Machine learning techniques, on the other hand, offer the potential to dynamically detect and respond to these threats based on the analysis of URL characteristics and behavioral patterns. This paper presents "Malware Link Detector," a proactive, machine learning-based system that not only blocks malicious URLs but also suggests safe, related alternatives to users when a threat is detected. By combining URL analysis with safe redirection, this system contributes to a safer and more user-friendly browsing experience. The structure of this paper is as follows: Section 2 reviews related work, Section 3 discusses the methodology and system architecture, Section 4 presents experimental results, and Section 5 discusses findings and future work.
2. Related Work Malicious URL detection has become an active area of research in cybersecurity, with traditional methods including blacklistbased and rule-based detection systems. Blacklists, like Google Safe Browsing, allow systems to quickly flag known malicious URLs but struggle with new, unlisted threats. Rule-based systems leverage predefined patterns or heuristics but often lack the adaptability needed for evolving attack methods. In recent years, machine learning has emerged as a promising solution. Models such as Random Forest, Support Vector Machine (SVM), and Gradient Boosting classifiers have been applied successfully to distinguish between safe and malicious URLs. Deep learning, as demonstrated in the DEPHIDES study [1], has also shown high accuracy in phishing detection, particularly with Convolutional Neural Networks (CNNs). However, many of these solutions stop the blocking malicious links without considering the user’s intent, creating an abrupt browsing experience. Malware Link Detector innovates by providing safe redirection, thereby supporting safer online behavior without interrupting the user’s browsing flow.
3. Methodology The methodology behind Malware Link Detector focuses on three main components: URL analysis and classification, user redirection with alternative links, and real-time processing for seamless integration into web applications.
© 2024, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 272