Non-bipolar evaluation and visualization of online text reviews

Abstract

Research in the field of Sentiment analysis is blowing day by day exponentially in the recent past and currently it is the much-acknowledged discipline. This is due to the outpouring users of internet and thus generated effusion of data in the form of reviews, comments, blogs, communications, etc. These are treasure trove of information needed to comprehend the fact-based opinions of diverse consumers. Elucidating those contents is propitious to various stakeholders as right sentiments can be gathered through it. But trading with such ginormous unstructured inputs embodies diverse challenges. These challenges trigger the raise in desideratum and thus summon the need to instigate pioneering logics to meet the same and to optimize the existing approaches. Sentiment analysis and opinion mining terms are used equivalently. This existed from ancient days. But the modes employed to perform this were different like in the form of surveys, elections, etc. It was carried out either by individuals or a group or an organization. Individuals used to consult friends or family before changing themselves from strangers to customers of any product or audience of any events. Groups or Organizations used to conduct surveys. But gradually most of the decisions were to be data-driven and thus appropriate decisions were to be made in short-run to bridge the gap between businesses and consumers and to gain the competitive advantage in market. So now this process was to be automated for mutual benefits by exploiting the escalation in technology. Various approaches are emerging every day to strengthen and ease the process. In the same direction, this paper addresses how sentiment analysis of online reviews can serve as suggestions to strangers and help them better in gaining additional insights to make well-judged decisions or choice by visualizing the analysis in the graph form.