Sentiment Analysis of Twitter Users on COVID-19 Vaccination Program in Indonesia using Support Vector Machine Algorithm

Abstract

The COVID-19 vaccination policy in Indonesia turns out to be both pros and cons. The government has to evaluate the underlying reason of why some people are against the policy, so that the vaccination program can run smoothly. Sentiment analysis as a way to see the polarity of opinion, makes it possible to classify positive, negative or neutral responses on Twitter regarding the vaccination policy. This study aims to determine the public's response to COVID-19 vaccination in Indonesia by examining word distribution and creating a Support Vector Machine (SVM) classification model. Sentiment analysis consists of several stages, namely data collection, data preprocessing, data weighting, data analysis, data sharing, classification modeling, hyperparameter tuning and model evaluation. The results of this study are a model with a relatively optimal performance in classifying sentiment with an accuracy, precision, recall and f1-score of 90%. The results of the sentiment analysis obtained are in the form of ideas, complaints, and suggestions for the COVID-19 vaccination.