Classification of Phishing Attacks Using Machine Learning Algorithms: A Systematic Literature Review

Abstract

Phishing attacks have been a major threat to cyber security since they take advantage of human vulnerabilities rather than system setbacks, making them difficult to detect. Phishing attacks always involve fraudulent websites designed to mimic legitimate websites to steal sensitive information from victims. This research paper provides a comprehensive literature review to recommend future research. This review paper examines previous papers' application of machine learning (ML) algorithms to phishing detection, focusing on how ML can be used to turn phishing attack problems into classification tasks. This research compared the commonly used ML algorithms like Decision Trees (DT), Random Forest (RF), Support Vector Machines (SVM), Naïve Bayes (NB), k-means Clustering, and Artificial Neural Networks (ANN), these algorithms were compared based on their performance, strengths, and weakness. Key findings reveal that SVM excels with high-dimensional data, RF handles large datasets efficiently, and DT offers simplicity but struggles with complex features. Algorithm performance depends on data and feature selection. This presents the need to develop hybrid or ensemble models to improve detection accuracy and reliability and contribute to stronger cybersecurity frameworks.