Sentiment Analysis of Twitter Data for COVID-19 Posts

Abstract

AbstractCOVID-19 (Coronavirus Disease of 2019) is now a genuine danger to the health of the world’s population. Vaccines capable of halting this pandemic have received much scientific and financial attention. These days expressing emotions over any social media platform is becoming a trend. This paper aims to study a technique known as sentiment analysis to learn about the sentiments of an individual. It takes a keyword as an input and will give you the polarity (positive or negative) as an output. Tweets can be positive for some and harmful for others. In order to get insight into how people perceive a particular tweet, the necessary steps are performed over a set of data. There is a tremendous amount of data available on the web for internet users, and much more data is being created as online technology continues to progress and flourish. The Internet has evolved into a medium for online education, idea exchange, and the expression of personal views. Twitter, Facebook, and Google + are among the most popular social networking sites because they allow users to share and express their opinions on various issues, engage with individuals from all backgrounds, and send messages worldwide. Sentiment analysis of Twitter data has gotten much attention. As a result, this survey focuses primarily on sentiment analysis of Twitter data, which helps analyze information in highly unstructured, varied, and neutral tweets. Sentiment analysis on Twitter has been explored as a whole and specific issues and uses. Some people favor getting vaccinations, while others are against them. This has led to much controversy among individuals. This study aims to develop a framework model for categorizing the sentiment and opinions expressed in tweets about the COVID-19 vaccinations. In order to better understand the people’s views on vaccines and focus their efforts on them, public health organizations need to identify such viewpoints. SVM and NB are two machine learning classification models employed here. Pre-processing approaches for unstructured tweets were also utilized.KeywordsMachine learningSentiment analysisTwitter