Wine Quality Prediction

A machine learning project that predicts wine quality based on various physicochemical properties. Using Wine Quality Dataset from https://www.kaggle.com/datasets/yasserh/wine-quality-dataset

Project Structure

wine-quality-prediction/
├── data/                   # Dataset directory
│   └── wine_quality.csv   # Wine dataset (to be added)
├── models/                 # Trained models
├── src/                    # Source code
│   ├── data_loader.py     # Data loading and preprocessing
│   ├── model_trainer.py   # Model training utilities
│   └── predictor.py       # Prediction utilities
├── notebooks/              # Jupyter notebooks
│   └── wine_analysis.ipynb # Data analysis notebook
├── train_model.py         # Main training script
└── requirements.txt       # Python dependencies

Setup

1. Install dependencies:

   pip install -r requirements.txt

2. Add your dataset:

   • Place your wine quality dataset as data/wine_quality.csv
   • Expected format: CSV with wine attributes and a 'quality' column
   • Common wine attributes: fixed_acidity, volatile_acidity, citric_acid, residual_sugar, chlorides, free_sulfur_dioxide, total_sulfur_dioxide, density, pH, sulphates, alcohol

Usage

Training Models

Run the main training script:

python train_model.py

This will:

• Load and preprocess the data
• Train multiple ML models (Random Forest, Gradient Boosting, Linear Regression, SVR)
• Compare model performance
• Save the best model to models/

Making Predictions

from src.predictor import WineQualityPredictor

# Initialize predictor with trained model
predictor = WineQualityPredictor('models/wine_quality_model.joblib')

# Predict single wine quality
wine_features = {
    'fixed_acidity': 7.4,
    'volatile_acidity': 0.7,
    'citric_acid': 0.0,
    'residual_sugar': 1.9,
    'chlorides': 0.076,
    'free_sulfur_dioxide': 11.0,
    'total_sulfur_dioxide': 34.0,
    'density': 0.9978,
    'pH': 3.51,
    'sulphates': 0.56,
    'alcohol': 9.4
}

predicted_quality = predictor.predict_single_wine(wine_features)
print(f"Predicted quality: {predicted_quality:.2f}")

Data Analysis

Open the Jupyter notebook for interactive analysis:

jupyter notebook notebooks/wine_analysis.ipynb

Models

The project includes several ML algorithms:

• Random Forest: Ensemble method, good baseline
• Gradient Boosting: Often performs well on tabular data
• Linear Regression: Simple interpretable model
• SVR: Support Vector Regression

Features

• Data Loading: Flexible CSV loading with preprocessing
• Model Training: Multiple algorithms with cross-validation
• Model Comparison: Automatic comparison of different models
• Feature Importance: Analysis of which features matter most
• Prediction Confidence: Uncertainty estimation for ensemble models
• Batch Prediction: Process multiple wines at once

Dataset

Expected wine attributes (typical wine quality dataset):

• fixed_acidity: Fixed acidity level
• volatile_acidity: Volatile acidity level
• citric_acid: Citric acid content
• residual_sugar: Residual sugar content
• chlorides: Chloride content
• free_sulfur_dioxide: Free sulfur dioxide
• total_sulfur_dioxide: Total sulfur dioxide
• density: Wine density
• pH: pH level
• sulphates: Sulphate content
• alcohol: Alcohol percentage
• quality: Target variable (wine quality score)