PhishGuard: A Smart Approach for Malicious URL Identification with ML

Abstract

With the rise of mobile device usage and increasing reliance on the internet, most real-world activities—banking, shopping, and communication—have transitioned online. While this digital shift enhances convenience, it also raises cybersecurity concerns, particularly phishing attacks. Phishing involves deceptive websites mimickinglegitimateonestostealsensitive user data like passwords and credit card numbers. Traditional security tools often fail to detect such sophisticated, zero-day attacks. This study suggests a machine learnings based phishing detections system that uses algorithms such as Random Forests, SVM, Decision Tree, Naïve Bayes and Neural Networks. These models analyze URL features to classifysitesaslegitimateorphishing.The system is deployed on a free, ad-free, non- profit website that also allows users to report suspicious URLs, enhancing accuracy over time. Tested on diverse datasets, the models achieved over 90% accuracy.Thisresearchhighlightsmachine learning’sroleineffectivelycombating phishing threats and strengthening cybersecuritydefenses.

Introduction

Phishing has become one of the most widespread and dangerous cybersecurity threats, especially with the growing use of the internet for banking, shopping, and online transactions. Phishing attacks are effective because they exploit human behavior, not just technical vulnerabilities.

• Attackers pretend to be trusted sources (e.g., banks, online retailers) to steal private information like passwords and account details.

• Even tech-savvy users can be deceived by realistic-looking fake websites.

• Annual global losses from phishing exceed $5 billion, with over $2 billion lost yearly in the U.S. alone.

2. Limitations of Traditional Methods

Traditional anti-phishing methods such as:

• Blacklists (known malicious URLs)

• Heuristic approaches (pattern-based detection)

...are no longer effective due to:

• Rapidly evolving attack methods

• Hidden or shortened URLs

• Auto-generated domain names

• High false-positive rates in heuristic methods

3. Role of Machine Learning (ML) in Phishing Detection

Machine Learning, a subset of Artificial Intelligence (AI), provides dynamic and pattern-based detection by learning from past phishing and legitimate websites.

Features commonly used:

• URL structure (length, presence of IP address, special characters)

• Domain age

• Use of HTTPS

• Similarity to known brands or websites

ML Algorithms Applied:

• Decision Trees

• Support Vector Machines (SVM)

• Naïve Bayes

• Neural Networks

???? Challenge: Lack of high-quality public datasets hinders effective model training.

4. Real-World Impact and Statistics

• Between 2020–2022, phishing attacks increased by 61%.

• 82% increase in fake domain usage.

• 55% of phishing websites mimicked popular brands (e.g., Amazon, Google, Facebook, Netflix).

5. Project Overview

The project introduces a user-friendly, web-based ML system to detect phishing websites. Key functionalities:

• Accepts a URL input and classifies it as phishing or legitimate

• Allows users to report suspicious URLs

• Displays prediction results and performance metrics like accuracy, recall, precision

Project Objectives:

• Create a fast, accurate phishing detection system using ML

• Use 32 features from URL, website content, and behavior

• Offer a simple dashboard with prediction results and charts

• Evaluate using metrics: Accuracy, Precision, Recall, False Positive Rate

6. Proposed System

A hybrid approach combining:

• Blacklist checking (fast filtering of known phishing sites)

• Random Forest ML model (detect new/unknown threats)

Features:

• Regularly updated blacklist

• High accuracy and efficiency

• Real-time analysis and predictions

7. Background Techniques

A. Phishing Website Discovery

1. URL Analysis:

   • Detects fake URLs using subtle changes (e.g., paypa1.com vs. paypal.com)

   • ML analyzes length, use of IPs, suspicious characters, subdomains, etc.

2. Content Analysis:

   • Examines text, images, links, and page structure

   • Looks for phishing signs like:

     • Urgent phrases (“Verify your account”)

     • Poor design

     • Brand logo misuse

     • Too many external links or scripts

B. Phishing Detection Approaches

References

[1] G. Karabatis and A. AlEroud, “Bypassing Detection of URL-Based Phishing Attacks Using Generative Adversarial Deep Neural Networks,” in Proc. 6th Int. Workshop on Security and Privacy Analytics, 2023. [2] M.Arivukarasi,A.Manju,R.Kaladevi, S. Hariharan, M. Mahasree, and A. B. Prasad, “Efficient Phishing Detection and Prevention Using SVM Algorithm,” in Proc. IEEE Int. Conf. on Communication Systems and Network Technologies (CSNT),Bhopal,India,pp.545–548,2023. [3] A. Mohamed, G. Özdemir, and S. Alrefaai, “Detecting Phishing Websites Using Machine Learning,” in Proc. Int. CongressonHuman-ComputerInteraction, Optimization and Robotic Applications (HORA), Ankara, Turkey, 2022. [4] B.Sabir,M.A.Babar,R.Gaire,andA. Abuadbba, “Reliability and Robustness Analysis of Machine Learning Based Phishing URL Detectors,” IEEE Transactions on Dependable and Secure Computing, 2022. [5] M.Almousa,T.Zhang,A.Sarraf Zadeh, and M. Anwar, “Phishing Website Detection: How Effective Are Deep Learning-Based Models and Hyperparameter Optimization?” Security and Privacy, vol. 5, no. 6, p. e256, 2022. [6] \"Deep Learning for Phishing Detection: Taxonomy, Current Challenges and Future Directions,\" IEEE Access, 2022, by N. Q. Do, A. Selamat, O.Krejcar, E. Herrera-Viedma, and H. Fujita. [7] M. D. Bhagwat, P. H. Patil, and T. S. Vishawanath, “A Methodical Overview on Detection, Identification and Proactive PreventionofPhishingWebsites,”in Proc. 3rd Int. Conf. on Intelligent Communication Technologies and Virtual Mobile Networks (ICICV), Tirunelveli, India, 2021. [8] Z.Alkhalil,C.Hewage,L.Nawaf,and I. Khan, “Phishing Attacks: Recent Comprehensive Study and a New Anatomy,”Computer Science Frontiers vol. 3, p. 6, 2021. [9] H. H. Chinaza Uchechukwu and J. Ding, \"A Survey of Machine Learning Techniques for Phishing Detection,\" IEEE Access, August 2020. [10] M. Korkmaz, O. K. Sahingoz, and B. Diri, “Detection of Phishing Websites by Using Machine Learning-Based URL Analysis,” in Proc. Int. Conf. on Computing,Communication and Networking Technologies (ICCCNT), Kharagpur, India, pp. 1–7, 2020. [11] \"Phishing Website Classification and Detection Using Machine Learning,\" with J. Kumar, A. Santhanavijayan, B. Janet, B. Rajendran, and B. S. Bindhumadhava, in Proc. Int. Conf. on Coimbatore, India: Computer Communication and Informatics (ICCCI), pp. 1–6, 2020. [12] OFS-NN: A Successful Phishing Website Detection Model Using Neural Networks and Optimal Feature Selection, IEEE Access, vol. 7, pp. 73271–73284, 2019; E. Zhu, Y. Chen, C. Ye, X. Li, and F. Liu. [13] IEEE Access, \"Detection of Phishing Websites Using Multidimensional Features Driven by Deep Learning,\" vol. 7, pp. 15196–15209, 2019. [14] Int. J. of Advanced Computer Science and Applications (IJACSA), vol. 10, no. 7, 2019, A. Kulkarni and L. L. Brown III, \"Detection of Phishing Websites Through Machine Learning.\" [15] According to R. Mahajan and I. Siddavatam, \"Detecting Phishing Websites Using Machine Learning Algorithms,\" International Journal of Computer Applications, vol. 181, no. 23, pp. 1–7, 2018.

Copyright

Copyright © 2025 Miss. Varalakshmi H, Mrs. Sharvani V. This 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.