Arabic Email Spam Detection Using Machine Learning and Deep Learning: A Comparative Study

Due to the relatively small figure of research conducted on Arabic email spam about the number of Arabic-speaking individuals using the internet, comes challenges related to Natural Language Processing in morphologically rich languages such as Arabic. The target goal of this research is to create a classification framework for spam email detection in Arabic. The recent dataset obtained from Kaggle contains phishing and legitimate emails written in Arabic. Term frequency - Inverse document frequency (TF-IDF) and word embeddings are some of the text preprocessing methods used in this research. Support Vector Machine (SVM), Naïve Bayes, and Random Forest machine learning models are used, along with Long Short-Term Memory (LSTM) deep learning methods. The SVM performed best in the comparative analysis, peaking at an ROC-AUC of 98% and F1 score of 99%. These results will aid in shifting Arabic spam detection and email security in the Arabic digital world forward, showcasing the positive role machine learning can play in this process.