0
- Volume 2, Issue 1 2024
By Fatima Siddiqui
10.20547/aibd.242102
Keywords: Product review Analysis, Information Retrieval, Sentiment Analysis, Roman Urdu, CFGs.
Sentiment analysis in Roman Urdu is increasingly crucial for enhancing consumer decision-making in diverse product domains. This paper addresses the challenge of extracting sentiment from product reviews written in Roman Urdu, leveraging Context-Free Grammar (CFG) to classify reviews into positive, negative, or neutral sentiments. Our study utilizes a dataset of online product reviews sourced from e-commerce platforms, focusing on the automation of sentiment classification. We propose a comprehensive methodology for sentiment extremity classification, demonstrating promising results through sentence-level analysis. This research contributes to advancing sentiment analysis in NLP, particularly in under-resourced languages like Roman Urdu, highlighting both methodological innovations and practical implications for e-commerce and beyond. Future work aims to further refine our approach and explore additional nuances in sentiment analysis for enhanced decision support.