Research on automatic sentiment analysis of text movie reviews with machine learning methods

Abstract:
In today’s world, vast amounts of information are available to us. For decades, scientists have sought to develop the most efficient methods of querying these data and extracting the information needed. In this survey, we describe the basics of sentiment analysis and discuss prominent work in natural language processing, focusing on the domain of movie reviews. We discuss methods that operate using lexicon-based approaches, conventional machine learning approaches, deep learning approaches, and hybrid approaches. Compared with previous papers, this article offers more comprehensive coverage, as it includes the latest methodologies in the field, e.g., the use of BERT in sentiment analysis of movie reviews. We also highlight the limitations of research to date and point to potential directions for future work.