A Survey on Unified Phishing and Email Spam Detection Systems Using Machine Learning and NLP

Phishing and spam are not going away. If anything, the tools available to attackers are becoming more sophisticated — generative AI has already made it easier to craft convincing, personalised phishing messages at scale, and LLM-assisted attacks can now dynamically tailor content to individual targets based on publicly available social data. The research community's response needs to match that sophistication. This survey traced the evolution of detection techniques from simple Bayesian filters through ensemble machine learning and into NLP-enriched and deep learning approaches. Algorithms including SVM [22], Naive Bayes, Random Forest [21], Logistic Regression, and XGBoost [23] each contribute something distinct, and the strongest systems today combine several of them. NLP tools — TF-IDF [17], word embeddings [24], and contextual models like BERT [6] — have transformed what it means to analyse email and webpage content. We reviewed 17 representative works spanning URL detection, email classification, semantic analysis, ensemble methods, and survey studies, summarised in Table 1. What this survey keeps returning to is the gap between individual technical advances and systems that actually work together. Isolated architectures, unexplainable decisions, and blind spots for zero-day threats [19] are the gaps that real attackers slip through. Unified frameworks that connect phishing detection to email classification, explain their reasoning in plain language, and stay current through live threat intelligence APIs [27, 28] represent the direction the field needs to move in. The building blocks are fully available — what is needed now is the architectural vision to assemble them into a coherent whole.

[1] C. Anilkumar, A. Karrothu, N. Sri Mouli, and C. Bhanu Tej, "Recognition and Processing of Phishing Emails Using NLP: A Survey," in Proc. Int. Conf. Computer Communication and Informatics (ICCCI), 2023. [2] A. R. More, O. B. Shendkar, K. G. Pathak, P. D. Sanap, S. N. Pachore, and P. William, "Survey on Framework for Governing Modern Cyber Attacks in Computer Network," in Proc. ICCAKM, 2023. [3] A. Jain and B. Gupta, "A Machine Learning Based Approach for Phishing Detection Using Hyperlinks Information," Journal of Ambient Intelligence and Humanized Computing, vol. 10, no. 5, pp. 2015–2028, 2019. [4] A. Blum, B. Wardman, T. Solorio, and G. Warner, "Lexical Feature Based Phishing URL Detection Using Online Learning," in Proc. ACM Workshop on AI and Security, pp. 54–60, 2010. [5] S. Garera, N. Provos, M. Chew, and A. Rubin, "A Framework for Detection and Measurement of Phishing Attacks," in Proc. ACM Workshop on Recurring Malcode, pp. 1–8, 2007. [6] J. Devlin, M. Chang, K. Lee, and K. Toutanova, "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding," in Proc. NAACL, 2019. [7] P. Verma, V. Goyal, and Y. Verdhan, "Phishing Website Detection Using Neural Network and Deep Learning," International Journal of Computer Applications, vol. 176, no. 39, pp. 1–6, 2020. [8] I. Fette, N. Sadeh, and A. Tomasic, "Learning to Detect Phishing Emails," in Proc. Int. Conf. World Wide Web, pp. 649–656, 2007. [9] J. Ma, L. Saul, S. Savage, and G. Voelker, "Beyond Blacklists: Learning to Detect Malicious Web Sites from Suspicious URLs," in Proc. ACM SIGKDD, pp. 1245–1254, 2009. [10] T. Peng, I. Harris, and Y. Sawa, "Detecting Phishing Attacks Using Natural Language Processing and Machine Learning," in Proc. IEEE 12th Int. Conf. Semantic Computing (ICSC), 2018.