A Review of Phishing Detection Techniques Using Machine Learning and Natural Language Processing

The goal of phishing attacks, which rank high among cyber risks, is to steal sensitive information (such as login passwords, financial data, and more) from unsuspecting people and companies. Blacklist, whitelist, rule, heuristic, and signature-based detection methods are no longer suitable due to the rising sophistication of phishing schemes. As a result, ML, DL, and NLP are rapidly becoming into effective instruments for detecting both well-known and brand-new forms of phishing. Background on phishing assaults, old ways of detecting them, and recent advances in intelligent phishing detection are all included in this article. It delves into several machine learning algorithms such as decision tree, LSTM, ANN, CNN, and RNN, as well as deep learning models like KNN, NB, and SVM. Also, natural language processing methods including word embedding, sentiment analysis, text categorization, and tokenization are explored with the aim of detecting phishing material. Presenting a comprehensive, precise, and up-to-the-minute phishing detection system is the ultimate goal of this study, which focuses on machine learning frameworks, performance indicators, present obstacles, and future research directions.