An Empirical Study on Sentimental Analysis using Deep Learning

Abstract

A study of a person's attitude in terms of using several unstructured texts is denoted as Sentimental analysis or opinion mining. Opinion mining or sentimental analysis distinguishes as the degree of polarity discover. The estimation of tweet review topics and a product is a high-grade sentimental analysis. Natural language understanding was essential for such data; many challenges were present in the natural language processing field for sentimental analysis. Nowadays, many pieces of research consider deep learning-based techniques for sentimental analysis in the natural language processing field. In this study, 25 papers were reviewed through deep learning-based approaches. Measures, as well as achievements attained by various methods, were simplified. The survey described the improvements and a limitation of each method as well as it regards the challenges and future potential research which is to acquire high accuracy and precision in sentimental analysis. Taxonomy represents the study gap and it elaborates on the various approaches.