User Review Analysis in Social Forums Based on Sentiment Cateloging

Authors

  • Jeyakumar Durairaj, Balamanigandan R, Selvanayaki S, S. John Justin Thangaraj

Abstract

Sentiment cataloging is a essential chore in sentiment identification and analysis, and its aim is to categories the sentiment or point of view found in the social webpages with the given feedback or command section text. The process of sentiment identification and classification we proposed not only follows the practical methods in subject-based text classification but also involves the sentiment analysis. Those bags of words are used in information retrieval. Machine learning approaches as well will not perform an advanced analysis on sentiment classification as on outdated theme based classification and categorization. We propose a simple yet efficient model, called Globalized User Review Analysis (GURA) by using the property of feedback sense with the sentiment classification of basic two opposite classes of labels, we proposed an algorithm with the data expansion technique first by creating sample sentiment toggled comments. The unique and transferred comments are then constructed in accordance with the one-to-one correlation. Thereafter, we enhance the dual training (DT) algorithm and a dual forecasting (DF) algorithm separately, in order to make use of the existing original samples and the stored switched samples in pairs. This model helps in systematic training and an automated statistical classifier that can be achieved with the estimated predictions. The overall polarity of the reviews can be viewed in the sentiment graph.

Downloads

Published

2020-05-12

Issue

Section

Articles