SMS Spam Detection Using Machine Learning

A machine learning project that classifies SMS messages as Ham (legitimate) or Spam using NLP techniques.

Project Overview

This project builds an SMS spam detection system using machine learning and natural language processing techniques. The goal is to classify messages as either Ham or Spam using text vectorization and classification models.

Features

• Text preprocessing and label encoding
• CountVectorizer and TF-IDF feature extraction
• Naive Bayes and Linear SVM model comparison
• Confusion matrix visualization
• Precision, recall, and F1-score evaluation
• Real-time custom message prediction

Dataset

The dataset contains 5,572 SMS messages with two classes: Ham and Spam.

• Total Entries: 5,572
• Classes: Ham, Spam
• Class Distribution: Imbalanced, with more Ham messages than Spam messages

Dataset Source

The dataset used in this project is the SMS Spam Collection dataset.

Installation

Install the required dependencies using:

pip install -r requirements.txt

How to Run

1. Clone this repository.
2. Open the notebook file in Jupyter Notebook or Google Colab.
3. Ensure spam.csv is in the same folder as the notebook.
4. Run all cells in order.

Results

Model Configuration	Accuracy
CountVectorizer + Naive Bayes	98.85%
TF-IDF + Naive Bayes	96.27%
CountVectorizer + SVM	98.64%
TF-IDF + SVM	99.07%

Conclusion

TF-IDF + SVM performed best on the test set. This project shows that combining strong text features with a linear classifier works very well for spam detection.

Notebook

• SMS_Spam_Detection_Using_ML.ipynb

Dataset File

• spam.csv