Emotion Mining of Indonesia Presidential Political Campaign 2019 using Twitter Data

Abstract

Abstract The presidential election campaign in Indonesia which was held in 2019 attracted the interest of the Indonesian people. People expressing their emotion through social media, one of which is Twitter. Supervised learning is used to classify user’s emotion of that issues. The goal of this research is extracting the emotional content in tweets. The interest is in whether the text is an expression of Twitter users’ about the presidential candidates and whether this online forum represents the election results. For this purpose, text mining techniques are performed and we will compare Naive Bayes Classifier and SVM in classification process. The result showed that Naive Bayes Classifier and SVM have good performance for classifiying text and each of classifier have not outperform the others. The result also showed that stemming step in text pre-processing do not give any significant result in model accuracy, because twitter data use informal languange like abbreviation or slang.