Klasifikasi Analisis Sentimen Terhadap Calon Presiden 2019 Pada Media Sosial Twitter Menggunakan Metode Algoritma Naïve Bayes

Abstract

The existence of Twitter has been widely used by various levels of society in recent years. The public's habit of posting tweets to evaluate the presidential candidates is one of the media in representing the public response to the presidential candidates. Therefore in this study an analysis of public sentiments towards the 2019 presidential candidates will be revealed through the Twitter social network. The analysis was carried out using a tweet classification that contained public sentiment towards the 2019 presidential nomination, namely jokowi and prabowo. The classification method used in this study is Naive Bayes Classification (NBC). NBC is used to get the classification of positive and negative responses to the twitter and get a preference value from the community towards the 2019 presidential candidates. The results of the jokowi data percentage test methods are 25%, 50%, 75%, and 100% of the amount of data from the training data yielding an accuracy of 64.67%, 70.57%, 87.56%, 97.50% and for the test results the percentage of Pre -owo data 25%, 50%, 75%, and 100% of the amount of data from the training data resulted in an accuracy of 64.57 %, 81.67%, 64.22%, 62.67%. And for the results of testing the positive response of the people on Twitter with a value of perference value of 53% for Jokowi and 48% for Prabowo. Therefore sentiment classification using the Naive Bayes classification method can be used to measure the public response to the performance of 2019 presidential candidates.

Keywords: Twitter, naive bayes, sentiment analysis