Analysing Emotional Context in Video Captions: A Sentiment-based Study

Sentiment analysis involves examining human emotions and different mental states. It's a method for figuring out if a sentiment is positive, negative, or neutral. The collection of compelling and thought-provoking speeches by experts and business executives on a range of topics is known as the TED Talks. Their popularity has grown because of their ability to communicate difficult concepts, offering insightful analysis, and motivation, and stimulating intellectual curiosity succinctly and compellingly. Effectively measuring audience mood becomes more challenging as TED Talks draw larger audiences from throughout the world. Sentiment analysis is essential for comprehending a range of responses considering this increase in audience, ensuring that TED Talks continue to have an impact and be relevant to the wide range of people they draw. Therefore, Understanding the feelings and thoughts that the speakers in public-speaking presentations express has grown more crucial as TED Talk videos have proliferated. Due to their poor information consistency and quality, little effort has been made to extract patterns from these videos, even though many of them have received a sizable number of user views. To identify patterns and trends in the emotional content of these lectures and to comprehend how presenters express their opinions, we perform sentiment analysis on those public-speaking presentations’ video captions using transcripts and captions. Data normalization methods were applied to remove noise from the data. We created a system utilizing VADER and BING algorithms to conduct classification on the TED dataset. This analysis has repercussions for comprehending how TED Talks affect audiences and what is the emotional impact of the video, whether it is favorable, unfavorable, or neutral, to improve the standard for the upcoming discussions.