Sentiment Analysis for Hindi Language

Technologies Used: Python, NLP, Transformers, Microsoft Multilingual-MiniLM-L12-H384
Description:
Developed a machine learning-based solution for sentiment analysis in the Hindi language. The project evaluates various pre-trained transformer models to determine their accuracy, precision, recall, and F1-score.
Key Features:
Explored and fine-tuned multiple pre-trained transformer models for Hindi text analysis.
Achieved the highest accuracy of 82% using the Microsoft Multilingual-MiniLM-L12-H384 model.
Conducted extensive evaluation to compare performance metrics across different models.
Outcome:
Published a research paper in IEEE showcasing the findings. The project is a significant contribution to Natural Language Processing (NLP) for low-resource languages.