🍷 Wine Quality Analysis and Prediction

This project performs exploratory data analysis (EDA) and builds classification and regression models to analyze wine characteristics and predict wine quality using various machine learning algorithms.

---

📚 Table of Contents

• Dataset
• Exploratory Data Analysis (EDA)
• Data Preprocessing
• Model Training and Evaluation
• Results
• Conclusion
• How to Run
• Dependencies

---

📂 Dataset

The dataset contains physicochemical properties and sensory quality ratings of red and white Portuguese "Vinho Verde" wines.

Each record includes attributes like:

• Fixed acidity, volatile acidity, citric acid
• Residual sugar, chlorides, free sulfur dioxide
• Density, pH, alcohol content
• Quality score (target)

---

📊 Exploratory Data Analysis (EDA)

EDA includes:

• Distribution plots of numerical features
• Correlation heatmaps
• Outlier detection
• Wine type comparison (red vs white)
• Quality class distribution

---

🧹 Data Preprocessing

• Imputation: Fill missing values using median
• Outlier Handling: Clip extreme values per wine type
• Feature Scaling: Normalize for distance-based models
• One-hot encode wine types (if needed)

---

🧠 Model Training and Evaluation

🔸 Classification Models

• Logistic Regression
• Support Vector Machine (SVM)
• Decision Tree
• Random Forest
• K-Nearest Neighbors (KNN)
• Gaussian Naive Bayes

Evaluation Metrics: Accuracy, Precision, Recall, F1-score

🔸 Regression Models

• Linear Regression
• Huber Regressor
• RANSAC Regressor
• Theil-Sen Regressor
• Decision Tree Regressor
• Random Forest Regressor
• Support Vector Regressor (SVR)
• KNN Regressor

Evaluation Metrics: Mean Squared Error (MSE), Root Mean Squared Error (RMSE)

---

📈 Results

✅ Classification Results

Model	Accuracy	Precision	Recall	F1-score
Logistic Regression	97.69%	97.84%	99.06%	98.45%
SVM	92.62%	92.26%	98.23%	95.15%
Decision Tree	98.38%	98.96%	98.85%	98.90%
✅ Random Forest	99.62%	99.58%	99.90%	99.74%
KNN	95.62%	96.69%	97.39%	97.04%
Gaussian Naive Bayes	97.15%	98.94%	97.18%	98.05%

---

✅ Regression Results

Model	MSE	RMSE
Linear Regression	0.5300	0.7280
Huber Regressor	0.5373	0.7330
RANSAC Regressor	0.7293	0.8540
Theil-Sen Regressor	0.5428	0.7368
Decision Tree Regressor	0.7069	0.8408
✅ Random Forest Regressor	0.3704	0.6086
SVR	0.6099	0.7809
KNN Regressor	0.6318	0.7948

---

🏁 Conclusion

• ✅ Random Forest Classifier achieved the highest classification performance.
• ✅ Random Forest Regressor outperformed all others in predicting quality ratings.
• Preprocessing and EDA significantly improved performance and interpretability.

---

▶️ How to Run

1. Clone the repository
2. Install dependencies
3. Run the notebook: Wine_prediction.ipynb

---

📦 Dependencies

Install required packages:

pip install pandas numpy matplotlib seaborn scikit-learn