Sentiment analysis using fuzzy naïve bayes classifier on covid-19

Abstract

Fuzzy Naive Bayes Classifier method has been widely applied for classification. The Fuzzy Naive Bayes method which consists of a combination of two methods including fuzzy logic and Naive Bayes is used to create a new system that is expected to be better. This research aim to find out the society's sentiments about COVID-19 in Indonesia and the use of the results of the Fuzzy Naive Bayes Classifier. The data of this research is obtained by scraping on Twitter in the period from January 1, 2020 to April 30, 2020. The classification method used in this research is the Fuzzy Naive Bayes Classifier method by applying the fuzzy membership function. In this research, sentiment analysis uses input data whose source is taken from tweets and the output data consists of sentiment data which is classified into three classes, namely positive class, negative class, and neutral class. In the distribution of training and testing data of 70%: 30%, the accuracy of the classification model using the confusion matrix is 83.1% based on 1199 tweet data consisting of 360 testing data and 839 training data. Also the presentation of each sentiment class was obtained which was dominated by positive sentiments, namely the positive class by 36.7%, the negative class by 35.0%, and the neutral class by 28.3%. Based on the results of the presentation, it can be concluded that there are still many people who have positive opinions or give positive responses to the presence of COVID-19 in Indonesia.