Comparative Analysis of Deep Learning, SVM, Random Forest, and XGBoost for Email Spam Detection: A Socio- Network Analysis Approach

The surge in email usage has led to an upsurge in spam emails, posing threats to both security and user experience. In this era marked by ever-evolving cyber threats, the task of effectively detecting email spam remains critical. This research presents a comprehensive comparative analysis offour well-known machine learning techniques: Deep Learning, Support Vector Machine (SVM), Random Forest, and XGBoost, specifically aimed at detecting email spam. The approach goes beyond analyzing email content alone; it also considers the social network characteristics of emails to enhance overall detection accuracy. This involves examining factors such as sender-receiver relationships, temporal patterns, and the community structure within email networks. To ensure the robustness and real-world relevance of our analysis, the research employs a diverse and extensive dataset containing both legitimate and spam emails for training and evaluation. By bridging the gap between cutting-edge machine learning and email security, this study contributes to the ongoing efforts to combat the growing menace of email spam.