Analysis of public opinion sentiment against COVID-19 in Indonesia on twitter using the k-nearest neighbor algorithm and decision tree

Abstract

COVID-19 has become an ongoing disease pandemic across the globe. The need for information makes social media such as twitter a place to exchange information. This tweet can be used to see public sentiment towards COVID-19 in Indonesia. Sentiment analysis classifies opinions from tweets that have been processed and classified into different sentiments, namely negative, neutral, or positive. The aim of this paper is to find the algorithm that has the best accuracy. The researcher proposes to compare the K-Nearest Neighbors (KNN) and decision tree algorithms to be used in the classification of sentiment data from tweets related to COVID-19 that took place in Indonesia. The results of the evaluation of performance metrics concluded that the decision tree algorithm has a higher level of accuracy than KNN. Decision tree produces accuracy = 0.765, error = 0.235, recall = 0.76, and precision = 0.767 which is better when compared to KNN which produces accuracy = 0.69, error = 0.31, recall = 0.66, and precision = 0.702.