Proposed Sentiment Analysis of COVID-19 Tweets using Machine Learning Algorithm

Social media coverage of COVID-19 has facilitated the analysis of pandemic sentiment through the application of various Machine Learning Algorithms. Twitter is a prominent a social media platform where individuals share tweets related to COVID-19, such as the count of positive cases or deaths, inducing a surge of apprehension and anxiety gripping individuals worldwide. This presents a chance for researchers and data analysts to acquire data intended for scholarly and research objectives. Primary objective of this study is to evaluate the accuracy of sentiment analysis of people’s perceptions regarding COVID-19 related news. We gathered Twitter data by focusing on hashtag keywords related to the coronavirus, COVID-19, new case, and deaths. The data was categorized into positive or negative categories utilized machine learning algorithms, including the multilayer perceptron, naive Bayes, random forest, and support vector machine, for the classification process. The algorithms were compared for accuracy using the datasets. All of the algorithms demonstrated a high degree of accuracy in their predictions, with the support vector machine algorithm achieving the highest classification accuracy of 99%. Consequently, the findings demonstrate that 70% of the sentiments exhibited positivity, while the remaining 30% conveyed negativity.