Petrolino/EmailSecurity

EmailSecurity: Multiclass Email Security Detection

EmailSecurity is a cybersecurity-focused email classification system that analyzes email subject and body text and classifies messages into four categories:

• —Ham
• —Spam
• —Phishing
• —Promotion

The application uses natural language processing and trained machine learning models to support practical email threat analysis.

Live Functionality

Users can enter an email subject and message body. The system returns:

• —Predicted email class
• —Confidence score
• —Risk level
• —Security recommendation
• —Basic text indicators
• —Class probability distribution

Classification Categories

--> Ham: Legitimate email communication --> Spam: Unwanted or low-value bulk email | --> Phishing: Potentially malicious email attempting to steal credentials, money, or sensitive informatioN. --> Promotion: Marketing, advertising, discount, campaign, or commercial email content.

Model Architecture

The system follows a two-layer classification design:

1. 1.Layer 1 separates legitimate emails from suspicious emails.
2. 2.Layer 2 classifies suspicious emails as spam, phishing, or promotion.

Author

Petrolino Grave do Espirito Santo Cybersecurity, Machine Learning, and Applied Research