Multilabel Datasets with LIFT Training System

A comprehensive system for training, evaluating, and deploying multilabel classifiers using the LIFT (Learning with Label-Specific Features) algorithm on various multilabel datasets.

🚀 Quick Start

Get started quickly with the interactive interface:

python quickstart.py

This will guide you through setup and provide access to all features.

📦 Available Datasets

The repository includes 10 popular multilabel datasets:

• birds - Bird species classification from audio features
• bookmarks - Web bookmark categorization
• Cal500 - Music emotion classification
• corel5k - Image annotation with Corel 5K dataset
• delicious - Social bookmarking tag prediction
• Emotions - Music emotion classification
• enron - Email classification
• genbase - Gene functional classification
• mediamill - Video semantic annotation
• yeast - Yeast protein functional classification

🛠️ Installation

1. Clone and setup:

   git clone <your-repo-url>
   cd Multillabel-Datasets
   git submodule update --init --recursive
   python setup.py

2. Quick start:

   python quickstart.py

📚 Available Scripts

🎯 Interactive Training

python multilabel_trainer.py --interactive

• User-friendly dataset selection
• Automatic feature/label detection
• Hyperparameter optimization with Bayesian search
• Comprehensive evaluation and reporting

🔮 Model Inference

python lift_inference.py --interactive

• Load trained models
• Make predictions on new data
• Batch inference from CSV files
• Model evaluation on test data

🔍 Dataset Explorer

python dataset_explorer.py --interactive

• Analyze dataset characteristics
• Generate statistical reports
• Compare multiple datasets
• Export detailed HTML/JSON reports

📊 Batch Runner

python batch_runner.py --comparison

• Train models on multiple datasets
• Compare optimization strategies
• Automated experiment logging

🏃 Simple Example

python run_lift_experiment.py --dataset yeast

• Basic LIFT training example
• Good for understanding the workflow

🔧 Command Line Usage

Train a specific dataset:

python multilabel_trainer.py --dataset yeast --optimize

Make predictions:

python lift_inference.py --model trained_models/yeast_model.pkl --data new_data.csv

Analyze all datasets:

python dataset_explorer.py --all

Run batch experiments:

python batch_runner.py --datasets yeast emotions birds

📊 Features

• Automatic Dataset Processing: Smart detection of features and labels
• Hyperparameter Optimization: Bayesian optimization for best performance
• Comprehensive Evaluation: Multiple multilabel metrics and per-label analysis
• Interactive Interface: User-friendly menu-driven operations
• Batch Processing: Train multiple datasets automatically
• Model Persistence: Save and load trained models
• Rich Reporting: HTML and JSON reports with visualizations
• Dataset Analysis: Statistical analysis and comparison tools

📁 Project Structure

Multilabel-Datasets/
├── quickstart.py              # Interactive quick start interface
├── multilabel_trainer.py      # Main training system
├── lift_inference.py          # Model inference and evaluation
├── dataset_explorer.py        # Dataset analysis tools
├── batch_runner.py            # Batch experiment runner
├── run_lift_experiment.py     # Simple example script
├── setup.py                   # Installation script
├── requirements.txt           # Python dependencies
├── USAGE_GUIDE.md             # Detailed documentation
├── *.zip                      # Dataset files
├── LIFT-MultiLabel-Learning-with-Label-Specific-Features/  # LIFT submodule
└── [Generated directories]
    ├── extracted_datasets/    # Extracted dataset files
    ├── trained_models/        # Saved models
    ├── reports/              # Training reports
    ├── dataset_reports/      # Dataset analysis reports
    └── predictions/          # Inference outputs

📈 Example Workflow

1. Explore datasets:

   python dataset_explorer.py --dataset yeast

2. Train a model:

   python multilabel_trainer.py --dataset yeast --optimize

3. Make predictions:

   python lift_inference.py --interactive

4. Compare multiple datasets:

   python batch_runner.py --comparison

🎯 Evaluation Metrics

The system provides comprehensive multilabel evaluation:

• Overall Metrics: Hamming Loss, Jaccard Score, F1 Scores (micro/macro/weighted)
• Per-Label Metrics: Precision, Recall, F1 Score, Support
• Advanced Analysis: Label cardinality, density, frequency distributions

📖 Documentation

For detailed usage instructions, see USAGE_GUIDE.md.

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

📄 License

This project follows the same license as the LIFT package (MIT).

---

Get started now: python quickstart.py 🚀