Sentiment classification and analysis of twitter data on distance learning using ML classifiers

Abstract

Online education comes in shades of grey. We analyze public view on the pandemic in regards with mental stress, interrupted power supply, affordability and access to internet, flexibility of schedule, reduction of long-distance commute, risk of covid and other such consequences of online learning and Government’s take on the need for inclusive education policies. In this paper, we qualitatively inspect the consequences of COVID-19 pandemic on education of the students. This study primarily focuses on the response of students of all age groups, educators, college professors, school teachers and also parents of young students towards the approach of distance learning or Online education in the past two years. We have taken two datasets, first being the Twitter dataset comprising of tweets from around the whole world and second, dataset which is specific to tweets from India. The data has been extracted from twitter with the aid of twitter API and then two sentiment analysis approaches have been implemented, first Machine learning classifiers namely, Naïve Bayes, SVM, Random Forest, Logistic Regression, KNN, XG-Boost and secondly, Lexicon Based algorithms, VADER and TEXTBLOB. Upon performing the said approaches, the maximum accuracy achieved is 94%.