Public Sentiment Analysis of Covid-19 Vaccination Drive in Pakistan

Abstract

COVID-19 brought chaos to the globe for the last three years and so, throwing the everyday routine into disarray and triggering the collapse of the global economy while resulting in more than 2 million fatalities. To prevent a pandemic, the whole globe combined their efforts to find a cure and help develop a vaccine. Vaccinations commenced in 2020, and some countries began their vaccination campaigns at the end of the year, while others waited eagerly for the vaccinations to be approved. In the rapidly changing Covid-19 situation, social media is full of a broad variety of both good and bad tales. Though a number of people are interested in COVID vaccination, however, some are worried about the adverse effects and are still nervous over the dreadful ideas and after effects causing confusion. To assist the government in determining the public’s reaction to the authorized COVID-19 vaccinations, it is crucial to ascertain the public’s sentiments and views. In this study, the main objective is to check the response of the Pakistani community to vaccination against COVID-19. Up till now, statistical analysis is performed, sentiment analysis, and basic machine learning-based predictive analysis. Sentiment analysis is performed by using machine learning and deep learning algorithms. In our research, it is our finding that the neural network models CNN and LSTM models perform better than as compared to other Machine learning models like Na ̈ıve Bayes, Random Forest, and Logistic Regression. Although our study also concluded that between Na ̈ıve Bayes, Random Forest, and Logistic Regression. The Logistic Regression performs better on a larger dataset. With the results of Na ̈ıve Bayes at 63%, Random Forest at 63%, and Logistic Regression at 66%. While the accuracy of neural network models is CNN 81.1% and of LSTM is 81.3%. So, LSTM model is best to classify the tweets into their categories.