Spam Detection – NLP

Project Description

This project, developed for ProfessionAI, focuses on analyzing and classifying incoming emails, with a particular emphasis on identifying SPAM messages and studying their content.

The workflow is structured into four main tasks:

1. SPAM/HAM Classification with Naive Bayes
   A Naive Bayes classifier was trained to distinguish between legitimate (HAM) and SPAM emails.

2. Topic Modeling on SPAM Emails
   Topics were extracted from SPAM emails to identify recurring themes and patterns.

3. Semantic Distance between Topics
   The semantic distance between SPAM topics was calculated to evaluate the diversity of unwanted content.

4. Organization Extraction from HAM Emails
   Organizations mentioned in legitimate emails were identified, providing useful insights for business intelligence.

Added Value:
This pipeline improves anti-spam filters, enables a deeper understanding of SPAM trends and content, enhances communication security, and enriches decision-making processes with insights extracted from legitimate emails.

Project Files

• spam-detector-project.ipynb

Dataset

• spam-dataset.csv

Author

Stefano Trovato
GitHub Profile
LinkedIn