An Intelligent Cybersecurity System for Phishing Attack Detection Using Machine Learning

Abstract

As our online presence grows, from online chats to online transactions and information sharing, phishing has turned out to be a significant concern in the world of cybersecurity. The term \"phishing\" describes a technique where hackers try to trick people into visiting fake websites and then persuade them to reveal their important information, like passwords or bank account numbers. The traditional \"blacklist-based\" technique, where a website\'s URL is checked against a list of known phishing websites, may not always be efficient in detecting new phishing websites. This project focuses on designing a machine learning-based phishing website detector that helps identify suspicious websites based on URL-based features. The machine learning algorithm will be trained on a dataset that consists of both phishing and legitimate websites. The features, such as URL length, domain name, presence of special characters, and security features, are extracted from the URLs and then used for classification. A simple web application, developed using the Flask web development library, can be used for this purpose. The results demonstrate that our system provides high accuracy with few false alarms. The approach is simple and flexible and can be easily incorporated into various cybersecurity tools in the real world. In summary, the approach is reliable and provides a solution for detecting phishing sites in the modern web environment.

Introduction

The paper presents a machine learning-based phishing detection system that identifies phishing and legitimate URLs through a lightweight, web-based framework. Traditional phishing detection methods such as blacklists and heuristic rules are ineffective against new (zero-day) phishing attacks, while many existing machine learning and deep learning approaches require high computational resources. To address these limitations, the proposed system balances high detection accuracy, low computational cost, and real-time performance.

The framework uses a structured dataset containing both phishing and legitimate URLs collected from public repositories. During preprocessing, duplicate and invalid URLs are removed, and important URL features—such as URL length, HTTPS usage, domain characteristics, special characters, IP addresses, and suspicious keywords—are extracted, normalized, and balanced using an 80:20 train-validation split. These features are then used to train a LightGBM-based binary classifier that predicts whether a URL is legitimate or phishing based on probability scores.

The proposed architecture consists of three main components: a URL feature extraction module, a machine learning classification module, and a Flask-based web interface. When a user submits a URL, the system automatically extracts its features, classifies it using the trained model, and displays the prediction in real time. By combining efficient feature engineering with lightweight machine learning, the proposed framework offers a practical, scalable, and accurate solution for modern phishing detection while remaining suitable for deployment in browsers, email systems, and network security applications.

Conclusion

This paper proposed a simple yet efficient phishing detector that utilizes machine learning techniques. The model is able to classify whether the given URL is a phishing or legitimate site by analyzing the basic features of the given URL. The results showed that the model achieved a total accuracy of 97%, where the model is equally effective in detecting phishing and legitimate sites, and the false positive rate is very low. The model is able to achieve such results because it is able to learn important features from the given URLs without adding complexity to the model. Moreover, the model is effective regardless of the dataset. However, the model is not perfect and has some limitations. Phishing sites that are similar to legitimate sites are difficult to detect, and the model only considers the static features of the given URLs. For future work, the model could be extended in the following ways: adding more features to the model to improve the accuracy of the results and testing the model with other datasets to further improve the accuracy of phishing URL detection. Future work will focus on incorporating real-time URL streaming data and advanced ensemble learning techniques to further enhance detection accuracy and adaptability.

References

[1] A. K. Jain and B. B. Gupta, “Phishing detection: Analysis of visual similarity-based approaches,” Security and Communication Networks, vol. 10, no. 8, pp. 1448–1463, 2017, doi: 10.1002/sec.1457. [2] R. Verma and A. Das, “What\'s in a URL: Fast feature extraction and malicious URL detection,” Proc. IEEE International Conference on Data Mining Workshops (ICDMW), 2017, pp. 914–921. [3] S. Marchal, J. François, R. State, and T. Engel, “PhishStorm: Detecting phishing with streaming analytics,” IEEE Transactions on Network and Service Management, vol. 11, no. 4, pp. 458–471, 2014. [4] M. Sahingoz, B. Buber, O. Demir, and B. Diri, “Machine learning-based phishing detection from URLs,” Expert Systems with Applications, vol. 117, pp. 345–357, 2019. [5] J. Ma, L. Saul, S. Savage, and G. Voelker, “Beyond blacklists: Learning to detect malicious web sites from suspicious URLs,” Proc. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2009. [6] H. Abutair and A. Belghith, “Using case-based reasoning for phishing detection,” Applied Soft Computing, vol. 13, no. 1, pp. 577–587, 2013. [7] M. Mohammad, F. Thabtah, and L. McCluskey, “Predicting phishing websites based on self-structuring neural networks,” Neural Computing and Applications, vol. 25, pp. 443–458, 2014. [8] Y. Rao and A. Pais, “Detection of phishing websites using machine learning approaches,” Procedia Computer Science, vol. 45, pp. 304–309, 2015. [9] B. B. Gupta, N. A. G. Arachchilage, and K. E. Psannis, “Defending against phishing attacks: Taxonomy of methods, current issues and future directions,” Telecommunication Systems, vol. 67, no. 2, pp. 247–267, 2018. [10] S. Sahoo, B. Liu, and S. C. H. Hoi, “Malicious URL detection using machine learning: A survey,” ACM Computing Surveys, vol. 50, no. 1, pp. 1–36, 2017. [11] A. Adebowale, K. Lwin, E. Sanchez, and M. Hossain, “Intelligent web phishing detection using machine learning,” Future Generation Computer Systems, vol. 108, pp. 425–435, 2020. [12] UCI Machine Learning Repository, “Phishing Websites Dataset,” University of California, Irvine, 2019. [13] A. Aljofey, Q. Jiang, H. Rasool, and X. Chen, “An effective detection approach for phishing websites using machine learning techniques,” IEEE Access, vol. 8, pp. 134–145, 2020. [14] S. Feng, R. Banerjee, and Y. Choi, “Syntactic feature-based phishing detection using machine learning,” Proc. IEEE International Conference on Communications (ICC), 2018, pp. 1–6, doi: 10.1109/ICC.2018.8422917. [15] T. Ahmad and U. A. Khan, “Phishing detection using URL-based features and machine learning algorithms,” IEEE Access, vol. 9, pp. 94752–94763, 2021, doi: 10.1109/ACCESS.2021.3094275. [16] S. Singh and P. Kumar, “URL-based phishing detection using machine learning classifiers,” Proc. International Conference on Computing, Communication, and Automation (ICCCA), 2020, pp. 1–5. [17] A. Abdelhamid, F. Thabtah, and H. Abdel-jaber, “Phishing detection: A recent intelligent machine learning comparison-based study,” IEEE Access, vol. 8, pp. 14110–14122, 2020, doi: 10.1109/ACCESS.2020.2965319. [18] N. Chiew, E. Chang, and K. S. Tan, “Utilizing hybrid features for phishing website detection,” Journal of Information Security and Applications, vol. 41, pp. 81–89, 2018. [19] A. Mishra and R. Gupta, “An efficient phishing detection model using machine learning techniques,” Procedia Computer Science, vol. 167, pp. 124–133, 2020. [20] M. Al-Ahmadi and H. Alharbi, “Phishing website detection using ensemble machine learning algorithms,” IEEE Access, vol. 9, pp. 150134–150146, 2021, doi: 10.1109/ACCESS.2021.3125984.

Copyright

Copyright © 2026 Raghupathi Hima Bindu, Pattapu Sai Deepthi, Vinukonda Geetha Sri, Velpula Beulah Rani, B Avinash. 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.