🚬 Smoking Prediction Web App using Linear Regression

This project uses machine learning to predict the likelihood of an individual being a smoker based on biometric and health signals. The application includes a trained model, a Flask web server, and a simple HTML frontend.

---

📌 Features

• 📊 Predicts smoking probability using 26 biometric and health metrics
• 🧠 Trained using scikit-learn Linear Regression
• 💾 Model serialized using pickle
• 🌐 Flask-powered web interface with user input form
• 🧼 Basic data preprocessing and encoding steps

---

🏗️ Project Structure

smoking-predictor/
├── app.py                 # Flask web server
├── model\_train.py         # ML training script
├── smoking.csv            # Dataset
├── mrs/
│   └── mr.pkl             # Saved trained model
├── templates/
│   └── index.html         # Web form interface
├── static/                # (Optional) CSS/JS assets
└── README.md              # Project documentation

---

🧠 Model Training

Run the training script to generate mr.pkl:

python model_train.py

This script:

• Loads and cleans the dataset
• Encodes categorical variables (gender, oral, tartar)
• Trains a Linear Regression model
• Saves the model in the mrs/ directory

---

🌐 Running the Web App

Install required libraries:

pip install -r requirements.txt

requirements.txt

pandas
numpy
scikit-learn
flask

Start the Flask server:

python app.py

Then open your browser at: http://127.0.0.1:5000

---

🖥️ Web Interface

The form takes in 26 numeric input values including:

• Gender (0 = Female, 1 = Male)
• Age, Height, Weight, Waist, Eyesight, Hearing
• Blood pressure, Sugar, Cholesterol levels, etc.
• Oral Health (Oral exam, Tartar, Dental Caries)

After submission, it displays the predicted probability of the person being a smoker.

📝 Note: All inputs are required and must be numerical.

---

🎯 Sample Output

Prediction of smoking is 0.83

---

💡 Future Improvements

• Switch from regression to classification (e.g., Logistic Regression or RandomForestClassifier)
• Add input validation and better UI/UX
• Deploy online using Render, Heroku, or Vercel
• Include real-world visuals or dashboard

---

📬 Contact

Contributions and questions are welcome! Raise an issue or submit a pull request.

---