<p>Hybrid recommender systems use advanced algorithms to learn from heterogeneous data sources and give consumers customized recommendations. User preferences (like ratings or reviews) and item content (like description or category) are two examples of the data that might be included. In earlier research on recommender systems, user feedback, or "ratings," was primarily used to construct user profiles and assess recommendation quality. Even if ratings are helpful, they might not give a complete picture of users' preferences. On the other hand, some feedback data—such as reviews and the emotions they convey—represent people and their preferences in a different or complementary way. Such information might highlight significant aspects of a user's profile that aren't always associated with user ratings; as a result, it may show a different aspect of the user. In this study, we provide a novel hybrid recommender system that analyzes heterogeneous data sources, such as user ratings and feelings gleaned from reviews, using sophisticated algorithms. Our method uses sentiment analysis to capture a more nuanced perspective of user preferences, in contrast to standard systems that build user profiles solely based on ratings. We employed sophisticated algorithms to provide recommendations for users who can incorporate extra data, such as the sentiment of the review. Our investigations revealed that in some cases (such as the music industry), the opinions expressed in user reviews do not always strongly correspond with the ratings. This suggests that emotion may represent a distinct facet of user preferences and serve as a substitute for other user feedback.</p>