📢 SMS Spam Detector

A web-based spam detection system built with React, Flask, and scikit-learn, capable of classifying SMS messages as spam or not spam in real-time.

🧠 Project Overview

This project was developed as part of our academic internship, where our goal was to build a reliable machine learning model to distinguish between spam and ham (non-spam) SMS messages using Natural Language Processing (NLP) techniques.

We used Python’s machine learning ecosystem to train, evaluate, and deploy a lightweight spam classifier accessible via a modern web interface.

---

👥 Team Members

• Faruk Khan
• Mridul Roy
• Mriganka Jyoti Deka
• Sanjeev Iqbal Ahmed
• Sourav Sharma

---

🚀 Features

✉️ For Users

• 📊 Instant SMS Analysis – Get real-time spam detection results
• 🔍 Spam Classification – Classifies messages as Spam ❌ or Not Spam ✅
• ⚡ Fast & Lightweight – Efficient predictions through pre-trained model
• 🌎 Online Access – Hosted using Render for seamless usage

---

🛠️ Tech Stack

Technology	Purpose	Version
React ⚛️	Frontend UI	18.2.0+
Flask 🐍	Backend API	2.3.2+
scikit-learn 📊	ML Model Training	1.3.0+
Pandas 📘	Data Handling	Latest
NumPy 🔢	Numerical Operations	Latest
Axios 🔗	API Integration	Latest
Gunicorn 🚀	Production WSGI	Latest
Tailwind CSS 🎨	UI Styling	3.3.0+

---

📥 Dataset & Model Info

• 📂 Dataset: SMS Spam Collection Dataset
• 🧠 ML Techniques:
  • Feature Extraction: TF-IDF Vectorizer
  • Algorithms: Multinomial Naïve Bayes, Logistic Regression
  • Final Model: MultinomialNB with TF-IDF
• ✅ Evaluation Accuracy: 97.6% on test dataset
• 📈 Evaluation Metrics Used: Accuracy, Precision, Recall, F1-score

---

🌍 Live Deployment

🔗 Try it live: https://sms-spam-detector-b15e.onrender.com

---

🛡 Future Enhancements

• 📊 Add confidence score for predictions
• 🌍 Multi-language spam detection
• 📈 Improve dataset with real-world examples
• 📥 Add email spam detection model