Analysis of Indonesian Society's Perceptions of the COVID-19 Vaccine in Youtube Comments Using Machine Learning Algorithms

Abstract

The coronavirus disease (COVID-19) was first identified in Wuhan, China, in December 2019. Based on the Indonesian COVID-19 Gugus Tugas report on July 14, 2021. This virus has infected 2.6 million people and has claimed the lives of 69,210 Indonesians. The spread of COVID-19 and the dangers it causes have prompted the government to take a policy to initiate the vaccination movement immediately. Currently, in Indonesia, the vaccination program is still running, but the response from the community is still pros and cons. This study aims to determine the results of Indonesian public opinion after the vaccination program runs by conducting sentiment analysis. The dataset used is Youtube commentary data, taken from the YouTube channel of the Indonesian Ministry of Health, the Presidential Secretariat, President Joko Widodo, and Najwa Shihab, using the keyword ‘Covid-19 Vaccine’. Comment data is retrieved from November 2020 to July 2021. The crawling of the data is done manually using the access token received from the YouTube API and the python programming language to extract the requested information data. Comment data is divided into three categories: positive, negative, and neutral. To find the best classification model, applying several machine learning algorithms, namely: K-nearest neighbor (KNN), Support Vector Machine (SVM), Multinomial Naive Bayes (MNB), Logistic Regression (LR), Random Forest (RF), Decision Tree (DT), Adaboost, and Stochastic gradient descent (SGD). The results of sentiment analysis research related to COVID-19 tend to get negative responses with sentiment measurements obtained more than 4.6 thousand positive comments (44.9%), more than 3.4 thousand positive comments (33.3%), and the rest more than 2.2 thousand. neutral comments (21.8%). The best accuracy value was obtained with the Linear Regression (LR) algorithm of 68%.