Sentiment Visualization of Covid-19 Vaccine Based On Naive Bayes Analysis

Abstract

COVID-19 is one of the topics that is being discussed intensively. The virus which was declared a global pandemic on March 11 by WHO caused around 2.09 million Indonesians to be infected with the COVID-19 virus. To overcome this, the government carried out a vaccination program. The data taken for this study is public opinion about the COVID-19 vaccine written on Twitter. The number of opinions written on Twitter requires classification according to the sentiments they have, whether they tend to be negative opinions or positive opinions using lexicon-based The idea of this research is to classify the covid vaccination dataset using the naive Bayes classifier method and visualization using word cloud. Crawling to obtain the dataset from Twitter, text pre-processing and labelling to determine the positive and negative classes, TFIDF feature extraction, data splitting with a percentage of 80% for train data and 20% for data testing, and finally classification using nave Bayes are the stages in this research The system's sentiment analysis research yielded significant results, the accuracy value is 73.1%, the precision value is 73% and the recall value is 83%.