Sentiview: AI-Driven Sentiment Analysis and Insight Extraction from YouTube Comments
Authors:-Associate Professor Dr.Deevi Hari Krishna, Teleprolu Sai Sri Harshini, Gudipati Pavani, Nagabhairu Varshitha
Abstract-In our research, we employ sentiment analysis to scrutinize public sentiment by processing comments on YouTube as our primary data source. Our goal is to identify a broad spectrum of sentiments, including positive, negative, and neutral, as well as to understand the underlying opinions and attitudes expressed in the text. We utilize comments on YouTube to gain insights into public viewpoints on various aspects of videos. Given that YouTube boasts over a billion unique users, it serves as a significant platform for this analysis. Users can voice their opinions through various means such as voting, rating, favoriting, sharing, and commenting on videos. We have designed a machine learning system grounded in Naïve-Bayes algorithms and assessed the accuracy of these classification algorithms using several metrics, including the F-score and Accuracy score. To boost our system’s performance, we have also incorporated numerous feature selection techniques. Our research adds value to the domain of text analytics, which involves scrutinizing unstructured data embedded in natural language text using various machine learning tools and methodologies. The insights we glean hold considerable potential for understanding user behavior and sentiment on platforms like YouTube.
