Machine Learning Analysis Toolkit

This project is a comprehensive desktop application developed in Python with Tkinter/PyQt5, designed for data exploration and comparison of various Machine Learning algorithms. It provides an interactive platform for analyzing datasets, particularly medical data such as the Pima Indians Diabetes dataset.

📸 Screenshots

Main application menu allowing users to load files and choose between supervised and unsupervised classification.

Analysis interface where users can select an algorithm and visualize results.

✨ Features

The application offers a complete pipeline from data preprocessing to model evaluation.

1. Data Analysis and Preprocessing

• Loading data files (ARFF format)
• Exploratory data analysis: descriptive statistics, distribution visualization (Boxplots, Scatter plots)
• Preprocessing: handling missing values, data normalization (Min-Max, Z-score)

2. Supervised Classification Algorithms

The application implements and compares the performance of the following algorithms:

• K-Nearest Neighbors (KNN)
• Gaussian Naive Bayes
• Decision Tree (with Gain Ratio and Gini Index)
• Neural Network (MLP)
• Support Vector Machine (SVM) with different kernels (RBF, Linear, Polynomial)
• Linear Regression adapted for classification

3. Unsupervised Learning

• Apriori Algorithm for association rule discovery and risk factor identification in data

4. Model Evaluation

• Comprehensive metrics: Precision, Recall, F1-Score, and Accuracy
• Visualization: Confusion matrix, metric evolution curves (for KNN), and decision boundary visualization (for Linear Regression and SVM)

🛠️ Technologies and Libraries

• Python 3
• Tkinter / PyQt5 for the graphical interface
• Pandas & Numpy for data manipulation
• Scikit-learn for models and metrics
• Matplotlib & Seaborn for visualizations
• SciPy for ARFF file loading
• PyDotPlus & Graphviz for decision tree visualization

🚀 How to Run

Prerequisites

Ensure you have Python 3.7+ installed on your system.

Installation Steps

1. Clone the repository:

   git clone https://github.com/yourusername/ML-Analysis-Toolkit-GUI.git
   cd ML-Analysis-Toolkit-GUI

2. (Recommended) Create a virtual environment:

   python -m venv venv
   # On Windows:
   .\venv\Scripts\activate
   # On macOS/Linux:
   source venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Launch the GUI application:

   python main_app.py

📂 Project Structure

ML-Analysis-Toolkit-GUI/
├── main_app.py              # Main application entry point
├── requirements.txt         # Python dependencies
├── src/
│   ├── algorithms/          # ML algorithm implementations
│   │   ├── knn.py
│   │   ├── naive_bayes.py
│   │   ├── decision_tree.py
│   │   ├── neural_network.py
│   │   ├── svm.py
│   │   └── linear_regression.py
│   ├── preprocessing/       # Data preprocessing modules
│   │   ├── data_loader.py
│   │   ├── data_cleaner.py
│   │   └── normalizer.py
│   ├── visualization/       # Visualization modules
│   │   ├── plots.py
│   │   └── metrics_viz.py
│   └── gui/                # GUI components
│       ├── main_window.py
│       ├── analysis_window.py
│       └── results_window.py
├── data/                   # Sample datasets
│   └── pima_diabetes.arff
├── docs/                   # Documentation and screenshots
│   ├── images/
│   └── user_guide.md
├── tests/                  # Unit tests
├── Rapport_Projet_ML.pdf   # Complete project report
└── README.md              # This file

🔧 Usage Guide

1. Loading Data

• Launch the application using python main_app.py
• Click "Load File" to import your ARFF dataset
• The application will automatically perform initial data analysis

2. Data Preprocessing

• Choose from various preprocessing options:
  • Handle missing values (mean, median, mode imputation)
  • Normalize data (Min-Max scaling, Z-score normalization)
  • Feature selection and engineering

3. Algorithm Selection

• Navigate to the analysis window
• Select from available algorithms:
  • Supervised Learning: KNN, Naive Bayes, Decision Tree, Neural Network, SVM, Linear Regression
  • Unsupervised Learning: Apriori algorithm for association rules

4. Model Training and Evaluation

• Configure algorithm parameters through the GUI
• Train the selected model on your dataset
• View comprehensive evaluation metrics and visualizations

5. Results Visualization

• Confusion matrices for classification performance
• ROC curves and precision-recall curves
• Decision boundary plots (where applicable)
• Feature importance analysis

🎯 Algorithm Implementations

Supervised Learning

K-Nearest Neighbors (KNN)

• Configurable K values
• Distance metrics (Euclidean, Manhattan, Minkowski)
• Performance visualization with varying K values

Support Vector Machine (SVM)

• Multiple kernel support (RBF, Linear, Polynomial)
• Hyperparameter tuning interface
• Decision boundary visualization

Neural Networks (MLP)

• Customizable network architecture
• Training progress monitoring
• Activation function selection

Decision Trees

• Splitting criteria (Gini Index, Gain Ratio)
• Tree pruning options
• Visual tree representation

Unsupervised Learning

Apriori Algorithm

• Association rule mining
• Support and confidence thresholds
• Risk factor identification in medical data

📊 Evaluation Metrics

The toolkit provides comprehensive model evaluation:

• Classification Metrics:

  • Accuracy
  • Precision
  • Recall
  • F1-Score
  • Specificity
  • Sensitivity

• Visualization Tools:

  • Confusion Matrix
  • ROC Curve
  • Precision-Recall Curve
  • Learning Curves
  • Feature Importance Plots

🔍 Data Analysis Features

Exploratory Data Analysis

• Statistical summaries
• Distribution analysis
• Correlation matrices
• Missing value detection
• Outlier identification

Data Preprocessing

• Missing value imputation
• Feature scaling and normalization
• Categorical variable encoding
• Feature selection techniques

🎨 GUI Features

• User-Friendly Interface: Intuitive design with clear navigation
• Real-Time Updates: Live visualization of training progress
• Interactive Plots: Clickable charts and graphs
• Export Functionality: Save results and visualizations
• Theme Support: Light and dark mode options

🚀 Advanced Features

• Model Comparison: Side-by-side algorithm performance comparison
• Cross-Validation: K-fold cross-validation implementation
• Hyperparameter Tuning: Grid search and random search options
• Batch Processing: Analyze multiple datasets simultaneously
• Export Reports: Generate comprehensive analysis reports

🧪 Testing

Run the test suite to ensure everything is working correctly:

# Run all tests
python -m pytest tests/

# Run specific test modules
python -m pytest tests/test_algorithms.py
python -m pytest tests/test_preprocessing.py

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Development Guidelines

• Follow PEP 8 style guidelines
• Add unit tests for new features
• Update documentation for API changes
• Use descriptive commit messages

🔧 Requirements

System Requirements

• Python 3.7 or higher
• Windows 10/macOS 10.14/Ubuntu 18.04 or later
• Minimum 4GB RAM recommended
• 1GB free disk space

Python Dependencies

tkinter>=8.6
pandas>=1.3.0
numpy>=1.21.0
scikit-learn>=1.0.0
matplotlib>=3.4.0
seaborn>=0.11.0
scipy>=1.7.0
pydotplus>=2.0.2
graphviz>=0.17

🐛 Troubleshooting

Common Issues

1. Graphviz Installation:

   # Windows (using conda)
   conda install graphviz
   
   # macOS
   brew install graphviz
   
   # Ubuntu
   sudo apt-get install graphviz

2. ARFF File Loading Issues:

   • Ensure your ARFF file follows the standard format
   • Check for special characters in attribute names
   • Verify data type consistency

3. Memory Issues with Large Datasets:

   • Consider data sampling for initial analysis
   • Use chunked processing for very large files
   • Monitor memory usage during processing

📌 Citation

If you use this project, please cite it as:

Ayoub Laib (2025), Machine Learning Analysis Toolkit, GitHub repository: https://github.com/aylaib/ML-Analysis-Toolkit-GUI/edit/main/README.md

🌟 Star History

If you find this project helpful, please consider giving it a star! ⭐

---

This comprehensive Machine Learning Analysis Toolkit demonstrates advanced skills in data science, software engineering, and GUI development, making it an excellent addition to any data science portfolio.