Publications

Abstract : Recently, the number of tweets on COVID-19 are increasing at an unprecedented rate by including positive, negative and neutral tweets. This diversified nature of tweets has attracted the researchers to perform sentiment analysis and analyze the varied emotions of a large public towards COVID-19. The traditional sentiment analysis techniques will only find out the polarity and classify it as either positive, negative or neutral tweets. As an advanced step, the proposed research work attempts to find the sentiment of tweets using Logistic Regression sentiment analysis, VADER sentiment analysis and BERT sentiment analysis. The proposed analysis methods are more sensitive to sentiment expressions in social media contexts, while it can be generalized on the basis of the domain. Even though 3 different algorithms are implemented, all the preprocessing and further steps excluding the sentiment analysis algorithm will remain identical. The identical processing steps will help to compare the proposed three different sentiment analysis algorithms. Furthermore, there are many useful applications with this proposed analysis, as this work obtains a public opinion for the government officials or even for the health officials and help them to work on the basis of the obtained results.