Covid-19 Sentiment Analysis Using Random Forest Classification

Abstract

The spread of the COVID-19 pandemic has reached a significant global scale, changing the dynamics of people's lives around the world. Social media platforms such as Twitter have become important channels for individuals to share experiences, voice opinions, and participate in discussions related to this pandemic. Sentiment analysis emerged as an important approach to reveal changes in people's attitudes and emotions in facing this challenge. This research involves analyzing sentiment during the COVID-19 pandemic to understand the feelings, attitudes, and views of the community after the peak phase of the pandemic. This study refers to previous findings which show that the Random Forest Algorithm provides the highest accuracy in this analysis. Through testing with the Random Forest Algorithm method, model accuracy testing is carried out using a confusion matrix and comparing test data and training data in a ratio of 80:20. Test results show that this model achieves an accuracy rate of 91%, providing a more comprehensive view of changes in public sentiment during the COVID-19 pandemic.