📩 SMS Spam Detection Web App

A simple web-based application built using Flask and Machine Learning to detect whether a given SMS message is Spam or Ham (Not Spam).

---

🚀 Features

• ✅ Predicts whether an SMS is spam or not
• ⚠️ Flash alert for spam messages
• 🧼 Clean text preprocessing using NLTK
• 🔍 Real-time text prediction using a trained ML model
• 🔄 Refresh/reset button
• 🖥️ Simple, clean Bootstrap UI
• 🔒 Lightweight and secure Flask backend

---

🧠 How It Works

1. Input: User enters an SMS message on the web interface.
2. Preprocessing: Message is lowercased, cleaned of punctuation, stopwords are removed, and stemming is applied.
3. Vectorization: Cleaned text is transformed using a trained TF-IDF vectorizer.
4. Prediction: The Multinomial Naive Bayes model predicts if the message is spam or ham.
5. Output: The prediction is shown with a flash message and result card.

---

📁 Folder Structure

spam_sms_detection_project/ ├── app/ │ ├── app.py # Flask application │ ├── templates/ │ │ └── index.html # Frontend ├── model/ │ ├── spam_classifier.pkl # Trained ML model │ └── tfidf_vectorizer.pkl # TF-IDF Vectorizer ├── spam.csv # Dataset (optional) ├── requirements.txt └── README.md

---

📦 Installation

Clone the repository:

git clone https://github.com/AnkithaMadhyastha/spam-detector-app.git cd spam-detector-app

---

📊 Dataset Used

The system uses the SMS Spam Collection Dataset, a labeled dataset of SMS messages marked as either spam or ham. It is widely used for NLP tasks and is included in the project as spam.csv.

Columns:

• label – spam or ham
• message – text of the SMS

---

💡 Sample Inputs

✅ Ham Message Example:

Hey! Just checking in. Are we still on for lunch today?

❌ Spam Message Example:

Congratulations! You’ve won a free vacation to Bahamas! Call now to claim.

---

🎯 Output Format

When a message is entered, the app returns:

• ✅ Ham – if the message is safe
• ❌ Spam – if it's a spam message

A flash message and label appear on the web page accordingly.

---

Snapsshot - output in this github project ,which is named has s1.png and s2.png

---

🔬 Model Training Info

The model was trained using:

• Algorithm: Multinomial Naive Bayes
• Vectorizer: TfidfVectorizer
• Text Cleanup:
  • Lowercasing
  • Removing punctuation
  • Removing stopwords
  • Applying stemming (PorterStemmer)

---

📦 License

Generate the spam_classifier.pkl and tfidf_vectorizer.pkl model training script for you too. This project is licensed under the MIT License.
You are free to use, modify, and distribute it with credit to the author.