Abstract:
Due to good consumer comments and reviews throughout the web, sentiment polarity detection has lately piqued the interest of NLP experts. The continued growth of e-commerce sites raises the purchasing rate of diverse items. People's interest in literature, for example, is fast increasing. Bangladesh already has a strong online marketing and e-commerce sector in this age of internet technology. Online product reviews, for example, have become a vital source of information for buyers making purchase decisions. It is believed that a person's best friend is a book. Books are essential to every person's existence because they provide knowledge about the outside world, help improve reading, writing and speaking abilities, and strengthen memory and intelligence. Our purpose is to rank Bangladeshi reviews and give accurate information about books and online bookstores in order to assist book lovers in purchasing the proper books and locating better online retailers. Using machine learning and natural language processing, this article demonstrates how to extract the sentiment polarity (positive or negative) from Bengali book reviews (NLP). We used five classification algorithm like: Multinomial Naïve Bayes(MNB), K-Nearest Neighbor (KNN), Random Forest Tree (RFT), Support Vector Classifier (SVC), and Stochastic Gradient Descent(SGD).
