Komparasi Algoritma Naïve Bayes Dan K-Nearest Neighbor Dalam Melihat Analisis Sentimen Terhadap Vaksinasi Covid-19

Abstract

Twitter is often used to deliver messages in the form of public opinion or opinion about the topic that is being reported. The government's policy to vaccinate has received various comments, ranging from praise, criticism, suggestions, and even hate speech. With so many twitter users who express their opinion, it can be used to find information. However, its use requires proper analysis, so that the resulting information can help many parties in making decisions or choices. Therefore, in this study, we tried to analyze sentiment on Covid-19 vaccination using the Naïve Bayes and K-Nearest Neighbor algorithms using the Cross Validation technique. The purpose of this study is to find out whether the Naïve Bayes and K-Nearest Neighbor algorithms in classifying produce optimal accuracy, to determine the sentiments of twitter users towards the Covid-19 vaccination and how much influence preprocessing has to measure accuracy on the classification. Based on the research that has been carried out, it can be concluded that the application of preprocessing for sentiment analysis on Covid-19 vaccination using the Naïve Bayes and K-Nearest Neighbor algorithms accompanied by the use of the Cross Validation technique got quite good results. The Naïve Bayes algorithm produces an accuracy of 77.62% and the K-Nearest Neighbor algorithm produces an accuracy of 76.43. Then for the positive response rate of the community to the Covid-19 vaccination, it was 55.63%.

Keywords: Comparison, Naïve Bayes, K-Nearest Neighbor, Sentiment Analysis,Vaccination, RapidMiner