ARCHIVES

This work presents a sentiment analysis system for the Assamese language, addressing the scarcity of Natural Language Processing (NLP) resources for low-resource Indian languages. The proposed system aims to automatically classify Assamese text into positive and negative sentiment categories using machine learning techniques. The framework applies VADER (Valence Aware Dictionary and sEntiment Reasoner) and Naive Bayes classifiers from the NLTK (Natural Language Toolkit) library, with the Google Translator API employed for text preprocessing and translation. The application is implemented using Streamlit to provide an interactive user interface and Ngrok for secure web-based deployment. The methodology involves translating Assamese text into English, followed by sentiment classification through VADER’s lexicon- based approach and Naive Bayes’ probabilistic model. VADER produces real-time sentiment scores—positive or negative, while Naive Bayes enhances accuracy through supervised learning on labeled data. Experimental evaluation demonstrates that VADER achieves rapid real-time performance, whereas Naive Bayes delivers higher classification accuracy after sufficient training. The combined approach ensures robust sentiment detection across diverse domains, including social media content, product reviews, and news articles. The system’s design emphasizes usability, accessibility, and reliability in multilingual sentiment analysis. This work contributes to the growing field of computational linguistics by providing a foundational sentiment analysis tool for Assamese text. Future research will focus on developing native Assamese sentiment lexicons, integrating deep learning models, expanding the training dataset, and incorporating contextual and domain-specific understanding to further improve performance.