Paraphrasing Tool for Low-Resource Languages

Overview

This project aims to develop a paraphrasing tool for low-resource languages, focusing on Marathi. The tool will transform complex sentences into simpler ones while maintaining their meaning, helping with language learning, NLP applications, and text simplification.

Features

• Sentence Validation: Ensures input sentences conform to the dataset.
• Paraphrasing Engine: Uses NLP techniques like tokenization, stemming, and synonym replacement.
• Evaluation System: Measures similarity between original and paraphrased sentences.
• Dataset Management: Supports dataset expansion for improved paraphrasing.

Technologies Used

• Programming Language: Python
• Libraries: NLTK, IndoWordNet, mahaNLP
• Platform: Google Colab, Spyder
• Machine Learning Models: Transformer-based models, LSTM

Installation & Usage

1. Clone the repository

   git clone https://github.com/your-repo/paraphrasing-tool.git
   cd paraphrasing-tool

2. Install dependencies

   pip install -r requirements.txt

3. Run the tool

   python main.py

System Requirements

• Hardware: Minimum 4GB RAM, Core i3 processor
• Software: Python 3.x, Google Colab (recommended)

Future Enhancements

• Implement lemmatization for better word replacement
• Expand dataset for improved paraphrasing
• Enhance semantic understanding using deep learning models

Contributors

• Pranita Barbade
• Akshada Malpure
• Anushka Pawar
• Reena Prasad