Spam Email Classifier

A machine learning project that detects whether an email/message is Spam or Not Spam using Natural Language Processing (NLP).

The system processes text messages, converts them into numerical features using TF-IDF vectorization, and predicts spam using a trained machine learning classifier.

---

Project Overview

Spam detection is a classic Natural Language Processing (NLP) problem. This project builds a complete spam detection pipeline that learns patterns from labeled messages and classifies new emails accordingly.

The pipeline includes:

1. Text cleaning
2. Tokenization
3. Stopword removal
4. TF-IDF feature extraction
5. Model training
6. Prediction on new messages

The trained model and vectorizer are saved and reused to allow instant predictions without retraining.

---

Technologies Used

• Python
• Pandas
• NLTK
• Scikit-learn
• TF-IDF Vectorization
• Logistic Regression

---

Dataset

The model is trained on the SMS Spam Collection Dataset, which contains 5,500+ labeled SMS messages categorized as:

• Spam
• Ham (Not Spam)

Each message is used to train the classifier to recognize spam patterns.

---

Machine Learning Pipeline

Raw Email
   ↓
Text Cleaning
   ↓
Tokenization
   ↓
Stopword Removal
   ↓
TF-IDF Vectorization
   ↓
Trained ML Model
   ↓
Spam / Not Spam Prediction

---

Installation

Clone the repository:

git clone https://github.com/sai161812/emaillclassifier.git
cd emailclassifier

Install dependencies:

pip install pandas nltk scikit-learn

Download required NLTK resources:

import nltk

nltk.download('punkt')
nltk.download('stopwords')

---

Model Persistence

The trained model and vectorizer are stored as serialized files:

spam_classifier.pkl
vectorizer.pkl

This allows the system to load the trained model instantly without retraining every time the program runs.

---

Example Predictions

Message	Prediction
"Claim your free prize now"	Spam
"Win a free lottery ticket today"	Spam
"Meeting scheduled at 3 PM tomorrow"	Not Spam
"Don't forget the project discussion"	Not Spam

---

Possible Improvements

Future enhancements may include:

• Spam probability score
• REST API using FastAPI
• Explainable AI for spam keyword detection
• Real-time email filtering integration
• Web interface for testing messages

---

Author

Sai Prasad B.Tech – Artificial Intelligence and Data Science

---