Classification of Book Review Sentiment in Bangla Language Using NLP and Machine Learning

Abstract

Books are said to be a person's best friend. Learning books is essential to acquire expertise. Community for book reading has been around for over a few thousand years. Since ancient civilization learns to write, knowledge is said to be the ancestors of books in tablets or walls or stones. The most up-to-date type of books is e-books, digital print or print on paper. In Bangladesh, even a few years ago, the people had to go to the library in person to gather books. Many of the advantages are simple and the internet bookshop has a penalty, i.e. the reader does not know the books or the book store itself. Book readers prefer to rely on feedback and ratings in order to prevent this. Our aim is to evaluate Bangladesh language reviews and to provide correct input on the books and online shop so that book readers can purchase correct books in order to read and get better online bookshops. About 1500 raw data have been obtained for training the system. In order to distinguish negative and constructive review results of previous users, we use Natural Language (NLP) processing. The data analyzers include a random forest, logistical regression, decision tree, random forest and some common algorithms like Subject Vector Machine (SVM)