📧 NLP Spam Detection System

👋 Overview

This project aims to build a robust NLP tool that:

1. Classifies emails as SPAM or NOT SPAM
2. Identifies key topics in spam emails and measures their semantic distance
3. Extracts organizations mentioned in non-spam emails

Leveraging libraries such as spaCy and NLTK, this system provides end-to-end solutions for text classification, topic modeling, semantic similarity, and named entity extraction.

---

🎯 Project Objectives

1. Train a Classifier for SPAM Identification

   • Use the provided dataset to train a machine learning model that accurately labels emails as SPAM or NOT SPAM.
   • Evaluate performance with metrics like accuracy, precision, recall, and F1-score.

2. Identify Main Topics in SPAM Emails

   • Perform topic modeling (e.g., Latent Dirichlet Allocation (LDA)) to uncover key themes in spam messages.

3. Calculate the Semantic Distance Between Topics

   • Measure how distinct each topic is using metrics like cosine similarity.
   • Assess diversity and overlap of the discovered themes.

4. Extract Organizations from NON-SPAM Emails

   • Apply Named Entity Recognition (NER) (using spaCy or NLTK) to detect and extract organization names in non-spam emails.

---

🏗️ Project Structure

1. Data Preprocessing

   • Clean and prepare the dataset for model training (e.g., removing noise, normalizing text).
   • Implement tokenization, stop-word removal, lemmatization, and stemming.

2. Classifier Training

   • Experiment with various algorithms (e.g., Naive Bayes, SVM, or neural networks).
   • Select and fine-tune the best model based on validation metrics.

3. Topic Modeling (SPAM Emails)

   • Use LDA to identify dominant topics in spam emails.
   • Visualize and interpret the most common themes.

4. Semantic Distance Computation

   • Implement methods (like cosine similarity) to measure how similar or different identified topics are.

5. Named Entity Recognition (NER)

   • Use spaCy or NLTK to detect and extract organization names from non-spam emails.

---

🔧 Libraries & Tools

• spaCy: Named Entity Recognition, tokenization, lemmatization
• NLTK: Tokenization, stop-word removal, text preprocessing

---

View my code on ipynb files! Happy coding! ✨