Sentiment Analysis for YouTube Cooking Recipes Videos Using User Comments

Nowadays, social media platforms have made many changes to human lifestyles. People can send messages, upload and view videos, and share their opinions and comments with others using these platforms. Among these platforms, YouTube is one of the best for sharing videos, especially cooking recipe videos. With the increasing popularity of these videos, sentiment analysis of user comments has become an important research area in the field of natural language processing (NLP). In the case of YouTube cooking recipe videos, sentiment analysis can provide valuable insights into how viewers perceive and respond to the videos and the recipes presented in them. Because when a user searches for a specific recipe, many videos will appear, and among those videos, some might be relevant and others might be irrelevant. Even though it has comments, reading comments and filtering relevant information remain difficult chores due to the vast amount of text. This research aims to develop a real-time system to provide recipe recommendations and classify comments into positive, neutral, and negative ways to facilitate the users' choice of the best and expected cooking recipe videos among those appearing on YouTube without spending too much time. Also content creators will benefit from the sentiment analysis results, which will provide them with insightful criticism on their cooking recipe videos that will help them better understand user preferences and improve their content. In this paper, the study propose a novel approach for Sentiment Analysis on Tanglish (Tamil and English) comments extracted from various Tamil YouTube channels using the YouTube API. The sentiments were classified using custom-built lexicon-based techniques, which include a comprehensive list of Tanglish words for cooking recipes. This approach addresses a research gap in sentiment analysis for non-standard languages, enabling accurate sentiment classification for cooking recipe videos.