📚 Roget's Thesaurus in the 21st Century

An investigation into how modern machine learning techniques align with Roget's classical thesaurus categorization.

✨ Features

1. Data Collection & Processing

   • Web scraping of Roget's Thesaurus
   • Hierarchical parsing of classes, divisions, sections, and terms
   • Custom preprocessing preserving linguistic nuances

2. Word Embeddings

   • OpenAI's text-embedding-3-large model
   • Parallel processing for efficient embedding generation
   • Chunked storage system for large-scale embeddings

3. Unsupervised Learning

   • Clustering analysis at both class and division/section levels
   • Comparison between discovered clusters and Roget's classification
   • Hungarian algorithm for optimal cluster matching

4. Supervised Classification

   • Two-level prediction models (class and division/section)
   • Neural networks, Random Forest and XGBoost implementations
   • Performance evaluation with multiple metrics

🛠️ Installation

1. Clone the repository:

git clone https://github.com/marsidmali/Roget-s-Thesaurus-in-the-21st-Century.git
cd Roget-s-Thesaurus-in-the-21st-Century

2. Create and activate a virtual environment:

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

3. Install dependencies:

pip install -r requirements.txt

4. Set up environment variables:
   • Copy .env.example to .env
   • Add your OpenAI API key

📁 Project Structure

Roget-s-Thesaurus-in-the-21st-Century/
│
├── notebooks/           # Jupyter notebooks
│   └── rogets_thesaurus_analysis.ipynb
│   └── assigment_3.ipynb
├── utils/               # Utility functions
│   └── thesaurus_parser.py
│   └── parallelized_embeddings_fetcher.py 
├── embeddings/          # Embeddings storage  
├── requirements.txt     # Dependencies
└── README.md            # Documentation

🚀 Usage

1. Launch Jupyter Notebook:

jupyter notebook

2. Open Roget's Thesaurus in the 21st Century.ipynb
3. Run all cells to perform the analysis

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.