Aspect-Based Sentiment Analysis on Candidate Character Traits in Indonesian Presidential Election

This paper is part of our research that was conducted during the campaign period of the Indonesian presidential election between 2018 and 2019. We collected millions of data from Twitter to predict the election results. In this research, we focus more on aspect-based sentiment analysis of the candidate character traits. We contribute a dataset that is annotated manually which contains the tweet texts in Bahasa Indonesia and their labels. The labels are the targets which are the candidates, the aspects which are the character traits, and the sentiments itself. We also present the comparison of machine learning algorithms' performances when classifying the dataset automatically. The results show that the Support Vector Machine algorithm performs better than the Naïve Bayes and the K-Nearest Neighbor.