A Survey on Challenges and Research Opportunities in Opinion Mining

Abstract

With huge development of internet and its usage, and with the advancement in latest mobile technology, internet usage and its users are essentially increasing day by day, and as a part of this, people are using this platform to express their feeling, opinion and reviews about their activity, feeling, frustration, place, person, sports, movie, politics, trolls, gossips, etc., i.e., today whatever happening around us, we are interested in sharing them either in our social media walls, or we want to express them, as a comments or reviews in many private blogs. Today even feedbacks were asked by many search engines like, “how about sharing your recent experience?”and consider Google itself, which do ask every users opinion about the recent activity they did. It may be a place they have visited, or shop where they did transaction, or new Gmail experience, or it may be anything. So many things happening around world are reached out in internet to people and also being shared very fast. Hence, opinion mining has become, a way to extract users feeling and analyze them for later usage. These extracted opinions are very important for manufacturers or the owner of an enterprise to modify their existing service, so as to upgrade and improve their quality and services that user wish for. Opinions are useful in all most every field. Opinions are not only useful but also misleading sometimes. Usually, public opinion consisting of both negative and positive feedback which plays a vital role, but while doing analysis we may face lot of new challenges. Challenges may be while identifying user views based on good or bad, worth or not worth, real or fake, etc. In today’s time, it’s not only text, but also emoji’s, hashtags and gif videos speaks about opinions. This paper consists of all those recent challenges that are faced with some examples, which may helpful to present issues and discussed with next state of text opinion mining process.