August 15, 2025Open Access

Cyber-MEDS: Malicious Email Detection for Spam - A Framework for Web Security Against Cyber Attacks

MSMuhammad Yasir ShabirUniversity of Azad Jammu and Kashmir NANour Ali Eid ALHomaidatAl-Hussein Bin Talal University AAAfshan AhmedUniversity of Azad Jammu and Kashmir

Key Points

Random Forest model achieved 95.11% accuracy, indicating high effectiveness in detecting malicious emails.
Evaluation measured precision at 95.89% and recall at 91.34% on the Spambase dataset, showcasing model performance.
Implementing machine learning techniques including Gaussian Naive Bayes and Gradient Boosting, improved detection rates.
Framework supports the development of scalable solutions for protecting users against cybercrime, emphasizing proactive measures.

Abstract

Email continues to be one of the most commonly exploited channels for launching cyberattacks, particularly through spam and phishing messages. These malicious emails pose serious threats to individuals and organizations by enabling data breaches,financial fraud, and malware infections. In this study, we propose a practical and efficient framework for detecting spam emails using a combination of machine learning algorithms. We evaluate models such as Random Forest, Gaussian Naive Bayes, Multi-Layer Perceptron, Gradient Boosting, and K-Nearest Neighbors on the publicly available Spambase dataset. Before training, we apply Min-Max scaling to normalize feature values and improve model performance. Among the tested models, Random Forest achieved the best results with an accuracy of 95.11\%, precision of 95.89\%, recall of 91.34\%, and F1-score of 93.56\%. Our framework shows that lightweight and well-tuned models can effectively identify malicious emails and serve as an early line of defense in email-based cybersecurity. This research contributes to the development of reliable and scalable solutions for protecting users against evolving cyber threats.

Ask AI

Helpful

Bookmark

View Full Paper