SMS Text Spam Prediction

A machine‑learning based project that detects whether an incoming SMS message is Spam or Ham (Not Spam).
This repository includes data preprocessing, feature engineering, model training, evaluation, and a prediction interface for classifying new text messages.

Overview

This project provides an end‑to‑end pipeline for SMS spam detection using classic NLP techniques and supervised learning models. It is designed for clarity, extensibility, and experimentation, enabling developers and researchers to quickly iterate and test different models.

Features

• Text preprocessing:
  • Cleaning, tokenization, stopword removal
  • Stemming or lemmatization
• Feature extraction:
  • Bag‑of‑Words (BoW)
  • TF‑IDF Vectorization
• Supervised ML models:
  • Naive Bayes
  • Logistic Regression
  • Support Vector Machines (SVM)
• Evaluation metrics:
  • Accuracy
  • Precision, Recall, F1‑Score
  • Confusion matrix
• Simple script or notebook to classify new SMS messages.

Installation

Clone the repository:

git clone https://github.com/FarooqMulla/SMSTextSpamPrediction.git
cd SMSTextSpamPrediction

Create a Virtual Environments (recommended)

python3 -m venv .venv
source .venv/bin/activate

Install Dependencies

pip install -r requirements.txt

Training Model

python src/train.py --data data/spam.csv

This will:

• preprocess text
• extract features
• train the classifier
• save the model and vectorizer under models/-0

Predict New Messages

Predict a message:

python src/predict.py --message "Congratulations! You've won a prize."

Predict from a file:

python src/predict.py --file data/sample_messages.txt

License

This project is licensed under the MIT License.