Comparative Classification Performance Evaluation of Various Deep Learning Techniques for Sentiment Analysis
Raj Gaurang Tiwari, Alok Misra, Neha Ujjwal
Abstract
In the discipline of text mining, the current subject of study is sentiment analysis. It detects textual subjectivity, feelings, and views. It is the use of text analytics and natural language processing tools to discover and extract subjective information from commonly utilized sources such as the web and microblogs. The prime aim of sentiment analysis is to study product and service evaluations and calculate sentiment ratings. The main issue is that the reviews are mainly unstructured, necessitating categorization or grouping to offer useful data for future use. Sentiment Analysis (SA) problem of Natural Language Processing (NLP) on text has been addressed in this research. With the help of neural networks trained on the “Movie Review Database,” the job of sentiment analysis from movie reviews has been completed. The ultimate accuracy of the trained network was 91 percent. The research adds to the body of knowledge by comparing the classification performance of several deep learning algorithms on a labeled corpus. This is intended to assist sentiment analysis specialists.