SENTIMENT ANALYSIS FOR E-BOOK REVIEWS ON AMAZON TO DETERMINE E-BOOK IMPACT RANK

Abstract

User-generated content platforms have changed the dynamics of the business environment and redefined how organizations and governments communicate with the public. Further, such platforms act as the primary means to measure customer satisfaction. Thus, those organizations need to analyze the content generated by their customer to extract their opinions then decide based on trustable information. Also, knowing user behavior and perception for a specific product is useful to customers in the decision-making process. In this thesis, a comparative study has been conducted to develop a model to measure customer satisfaction on Amazon e-book products by applying natural language processing (NLP), machine learning, deep learning, and text mining techniques on costumers reviews. This thesis will study the possibility of generating a rating based on sentiment analysis of each product instead of rating-based stars, which is already applied to the Amazon e-book rating system.