<p>The given paper introduces the hybrid recommendation approach (HRA) which combines sentiment analysis with content and collaborative filtering. The framework utilizes text-BERT (a pre-trained transformer model) and fine-tunes it to classify textual reviews for sentiment polarity scores. The textual reviews or input dataset is subjected to feature selection using Binary Horse Herd Optimization Algorithm (BHHOA). These sentiment insights are then combined with the collaborative filtering (CF), which deals with user-item interactions, and the content-based filtering (CBF), which utilizes item attributes to give full details and a personalized recommendation. In view of surpassing the limitations of traditional systems, the proposed framework provides accurate recommendation solutions for overcoming data sparsity and improves the comprehension of contexts. This approach discusses the possibility of adopting sentiment analysis procedure alongside novel hybrid recommendation approaches for improving user satisfaction and decision making in e-commerce systems.</p>