Sentiment Analysis Approaches and Applications - A Review

Abstract

With the advent of smartphones and the ease of access to the internet, people are mainly interested in sending textual messages through social media platforms. In many cases, customers would like to review the services provided by different providers in order to express satisfaction or dissatisfaction. The sentiments of users make a huge difference in the success of any business idea in the present digital age. As there are many competitors in every field of technology, health, and education, people would selectively want to use the resources that have positive opinions about them from the user community in the online reviews. There are different techniques to effectively estimate the user reviews, whether they are for or against a particular concept or the product. There are different techniques, like lexicon-based techniques, machine learning-based techniques, and deep learning-based techniques which are used to analyse the sentiments of the users’ reviews in order to improve user expectations. Lexicon-based techniques have many challenges, like the wrong interpretation of the meanings of the words and giving wrong sentiment scores to the words used by ignoring the grammatical constraints in the user reviews. There are many machine learning algorithms, like Logistic regression (LR), and Support Vector Machines (SVM) which can overcome the shortcomings of lexicon-based sentiment analysis models and could be used in various spheres of applications. The manuscript presents a detailed study in this regard.