Phishing emails continue to be an important threat to cybersecurity, necessitating the use of robust detection systems. On a large-scale dataset, this paper provides a holistic methodology for detecting phishing emails that combines NLP approaches with ensemble learning methods. AdaBoost, XGBoost, Gradient Boosting Machine (GBM), Light Gradient Boosting Machine (LGBM), CatBoost, Extra Trees, and Random Forest are among the ensemble approaches, along with the Stacking and Majority Voting procedures. The experimental results reveal that the Stacking ensemble obtained amazing performance, consisting of 98.89% accuracy, precision, recall, and F-measure, with unusually low FPR and FNR of 0.01 for each example. Similarly, the Majority Voting ensemble obtained remarkable results with an accuracy of 98.56%, precision, recall, and F-measure of 98.56%, FPR as 0.02, and FNR as 0.01. These findings highlight the capabilities of modern ensemble approaches to successfully detect phishing emails with high accuracy and low error rates. Combining natural language processing for feature extraction with complex ensemble models offers a viable method for combating phishing attacks in real-world applications.
SalloumSGaberTVaderaS, et al. A systematic literature review on phishing email detection using natural language processing techniques. IEEE Access2022; 10: 65703–65727.
4.
RashedSOzcanC. A comprehensive review of machine and deep learning approaches for cyber security phishing email detection. Al-Iraqia J Sci Eng Res2024; 3: 1–12.
5.
DoNQSelamatAKrejcarO, et al. Deep learning for phishing detection: taxonomy, current challenges and future directions. IEEE Access2022; 10: 36429–36463.
6.
ThakurKAliMLObaidatMA, et al. A systematic review on deep-learning-based phishing email detection. Electronics2023; 12: 4545.
7.
AlkhalilZHewageCNawafL, et al. Phishing attacks: a recent comprehensive study and a new anatomy. Front Comput Sci2021; 3: 563060.
8.
AbroshanHDevosJPoelsG, et al. Phishing happens beyond technology: the effects of human behaviors and demographics on each step of a phishing process. IEEE Access2021; 9: 44928–44949.
9.
CatalCGirayGTekinerdoganB, et al. Applications of deep learning for phishing detection: a systematic literature review. Knowl Inf Syst2022; 64: 1457–1500.
10.
Jáñez-MartinoFAlaiz-RodríguezRGonzález-CastroV, et al. A review of spam email detection: analysis of spammer strategies and the dataset shift problem. Artif Intell Rev2023; 56: 1145–1173.
11.
BasitAZafarMLiuX, et al. A comprehensive survey of AI-enabled phishing attacks detection techniques. Telecommun Syst2021; 76: 139–154.
12.
JampenDG’urGSutterT, et al. Don’t click: towards an effective anti-phishing training. A comparative literature review. Hum-cent Comput Inform Sci2020; 10: 33.
13.
Das GupttaSShahriarKTAlqahtaniH, et al. Modeling hybrid feature-based phishing websites detection using machine learning techniques. Ann Data Sci2024; 11: 217–242.
14.
ButtUAAminRAldabbasH, et al. Cloud-based email phishing attack using machine and deep learning algorithm. Complex Intell Syst2023; 9: 3043–3070.
GhalechyanHIsrayelyanEArakelyanA, et al. Phishing URL detection with neural networks: an empirical study. Sci Rep2024; 14: 25134.
18.
SonowalG. Phishing email detection based on binary search feature selection. SN Comput Sci2020; 1: 191.
19.
Ujah-OgbuaguBCAkandeONOgbujuE. A hybrid deep learning technique for spoofing website URL detection in real-time applications. J Elect Syst Inform Technol2024; 11: 7.
20.
AtawnehSAljehaniH. Phishing email detection model using deep learning. Electronics2023; 12: 4261.
21.
BaguiSNandiDBaguiS, et al. Machine learning and deep learning for phishing email classification using one-hot encoding. J Comput Sci2021; 17: 610–623.
22.
AlsuwaylimiAA. Enhancing arabic phishing email detection: a hybrid machine learning based on genetic algorithm feature selection. Int J Adv Comput Sci Appl2024; 15: 419–438.
23.
RabbiMFChampaAIZibranMF. Phishy? Detecting phishing emails using machine learning and natural language processing. In: Software engineering and management: theory and application: Volume 16, 2024, pp.119–137. Springer, Cham.
24.
JamalSWimmerHSarkerIH. An improved transformer–based model for detecting phishing, spam and ham emails: a large language model approach. Sec Privacy2024; 7: e402.
25.
AltwaijryNAl-TuraikiIAlotaibiR, et al. Advancing phishing email detection: a comparative study of deep learning models. Sensors2024; 24: 2077.
26.
SomeshaMPaisAR. Classification of phishing email using word embedding and machine learning techniques. J Cyber Secur Mobility2022; 11: 279–320.
27.
YasinAAbuhasanA. An intelligent classification model for phishing email detection. arXiv preprint. 2016. Available from: https://arxiv.org/abs/1608.02196.
28.
Dharmaraj PatilRWaghRBPunjabiVD, et al. Enhanced phishing URLs detection using feature selection and machine learning approaches. Int J Wireless Microw Technol2024; 14: 48–67.
29.
PatilDR. A framework for malicious domain names detection using feature selection and majority voting approach. Informatica2024; 48
30.
PatilDRPattewarTMPunjabiVD, et al. Detecting fake social media profiles using the majority voting approach. EAI Endorsed Trans Scalable Inf Syst2024; 11: 1–18.
31.
PatilDPattewarTPardeshiS, et al. Learning to detect phishing web pages using lexical and string complexity analysis. EAI Endorsed Trans Scalable Inf Syst2022; 10: 1–13.
ZhuCTangJLiH, et al. A Unified tagging approach to text normalization. In: Proceedings of the 45th Annual meeting of the association for computational linguistics, 2007, pp.688–95.
34.
M’ullerTCotterellRFraserA, et al. Joint lemmatization and morphological tagging with lemming. In: Proceedings of the 2015 conference on empirical methods in natural language processing, 2015, pp.2268–74.
35.
WuHCLukRWPWongKF, et al. Interpreting TF-IDF term weights as making relevance decisions. ACM Trans Inf Syst2008; 26: 1.
36.
BeelJBreitingerC. Evaluating the CC-IDF citation-weighting scheme – How effectively can ’Inverse Document Frequency’ (IDF) be applied to references? In: Proceedings of the 12th iConference, 2017. Available from: https://www.ideals.illinois.edu/handle/2142/97511.
37.
GoldbergYLevyO. word2vec Explained: Deriving Mikolov et al.’s Negative-Sampling Word-Embedding Method. arXiv. 2014. Available from: https://arxiv.org/abs/1402.3722.
FreundYSchapireRE. A decision-theoretic generalization of on-line learning and an application to boosting. In: Lecture notes in computer science, 1995, pp.23–37. Springer.
44.
ChenTGuestrinC. XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016.
45.
FriedmanJH. Greedy function approximation: A gradient boosting machine. Ann Stat2001; 29: 1189–1232.
46.
KeGMengQFinleyT, et al. LightGBM: a highly efficient gradient boosting decision tree. In: Proceedings of the 31st conference on neural information processing systems, 2017, pp.3146–56.
47.
DorogushAVErshovVGulinA. CatBoost: a high-performance gradient boosting library. In: Proceedings of the 2018 data mining and knowledge discovery conference, 2018.
SeewaldAK. How to make stacking better and faster while also taking care of an unknown weakness. In: Proceedings of the 19th international conference on machine learning, 2002, pp.554–61.
51.
SillJTakacsGMackeyL, et al. Feature-weighted linear stacking. In: Advances in neural information processing systems, Vol. 22, 2009.
52.
LittlestoneNWarmuthM. The weighted majority algorithm. Inf Comput1994; 108: 212–261.
53.
LittlestoneNWarmuthM. Weighted Majority Algorithm. In: Proceedings of the IEEE symposium on foundations of computer science, 1989.
