https://github.com/humayun-raza-030/wine-quality-feature-selection-and-model-optimization

https://github.com/humayun-raza-030/wine-quality-feature-selection-and-model-optimization

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/humayun-raza-030/wine-quality-feature-selection-and-model-optimization
• Owner: humayun-raza-030
• Created: 2025-02-23T07:01:33.000Z (3 months ago)
• Default Branch: main
• Last Pushed: 2025-02-23T07:06:04.000Z (3 months ago)
• Last Synced: 2025-02-23T08:18:43.020Z (3 months ago)
• Language: Jupyter Notebook
• Size: 233 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Wine-Quality-Feature-Selection-and-Model-Optimization

## Overview

This project focuses on **feature selection techniques** to enhance machine learning model performance on the **White Wine Quality dataset** from the **UCI Machine Learning Repository**. The dataset contains 11 chemical properties of white wine, and our goal is to identify the most relevant features that influence wine quality ratings.

## Dataset

- **Name**: White Wine Quality

- **Source**: [UCI Machine Learning Repository](https://archive.ics.uci.edu/ml/datasets/Wine+Quality)

- **Description**: The dataset consists of 11 input features (e.g., acidity, sugar, pH, etc.) and 1 target variable (wine quality, rated from 0 to 10).

## Tasks Performed

### 1. Data Preprocessing

- Checked for missing values.

- Analyzed the distribution of the target variable.

- Normalized the dataset (if necessary).

- Split the dataset into **training (80%)** and **testing (20%)** sets.

### 2. Feature Selection Techniques

#### **Filter Methods**

- **Low Variance Filter**: Removed features with very low variance.

- **Pearson’s Correlation**:

  - Computed correlation coefficients.

  - Visualized correlations using a heatmap.

  - Removed features with low correlation with wine quality.

  - Used **Variance Inflation Factor (VIF)** to detect multicollinearity.

- **Mutual Information**:

  - Computed mutual information scores.

  - Selected the top 5 most informative features.

#### **Wrapper Methods**

- **Forward Feature Selection**: Iteratively added features until model performance stopped improving.

- **Backward Feature Elimination**: Started with all features and iteratively removed the least important ones.

#### **Embedded Methods**

- **Lasso Regression (L1 Regularization)**: Selected features based on non-zero coefficients in Lasso regression.

### 3. Performance Comparison

- Compared the number of features selected by each method.

- Evaluated model performance using:

  - **Accuracy**

  - **Precision**

  - **Recall**

  - **F1-score**

- Identified the most effective feature selection technique.

## Results

- The best feature selection method was **Forward Selection**, as it provided the highest model accuracy while minimizing feature redundancy.

- The final model achieved an accuracy of **52.14%**.

## How to Use

1. Clone this repository:

   ```bash

   git clone https://github.com/humayun-raza-030/wine-quality-feature-selection.git

   ```