Jurnal Elektronik Ilmu Komputer Udayana p-ISSN: 2301-5373 Volume 9, No 3. February 2021 e-ISSN: 2654-5101 443 Sentiment Analysis Of Tribal, Religion, And Race With LIWC Prasetyo Adi Utomo a1 , AAIN Eka Karyawati a2 a Informatics Department, Faculty of Math and Science, Udayana University South Kuta, Badung, Bali, Indonesia 1 pras.au404@email.com 2 eka.karyawati@unud.ac.id Abstract During this pandemic, social media has become a major need as a means of communication. One of the social media used is Twitter by using messages referred to as tweets. In Indonesia itself, there are various tribes, religions, and races in their society so the use of these names is also become commonly used. However, sometimes, the use of the name is followed by negative sentiment that used to insult and aimed at an individual or group. To filter that kind of tweets, a sentiment analysis was performed with LIWC method that divides tweets into 3 classes of positive, neutral, and negative. From the sentiment analysis that has been performed, the average score for precision is 69.62%, recall is 70%, and f-measure is 69.81%. Keywords: Sentiment Analysis, Tweet, LIWC, Indonesia, Religion 1. Introduction The need for social media has become part of Indonesian society. Moreover, in present that the current Covid-19 pandemic happen where social media are used as the main means to communicate due to social restrictions to prevent the spread of viruses. One of the social media that used is the Twitter where users communicate using tweet as a message. Messages in tweets can contain such as congratulations messages, event descriptions, or can be issues and opinions about a person, politics, or government regulations. In Indonesia, there are many tribes, religions, and races where people often use the name of a tribe, religion, or race to indicate something like to congratulate someone or something else. In using tweets, people also often use those words. However, there are also tweets that use the name of a tribe, religion, or race and contain negative sentiment or bad words to insult an individual or a group. The use of such tweets can lead to fights between tribes, religions, or races. To prevent this, it is necessary to select or filter tweets that contain an insult or bad wording of tribal, religion, or race. To find out if a tweet contains a tribe, religion, or race and know the sentiment of the tweet whether it is positive, neutral, or negative can be done by applying sentiment analysis to the tweet. Sentiment analysis is a field of science that analyzes opinions, attitudes, evaluations, and assessments of an event, topic, organization, or individual [3]. In sentiment analysis itself, approaches that can be used are machine learning-based and lexicon-based approach. The example of machine learning-based is sentiment analysis using Naïve Bayes method which have been carried out by [7]. The research conduct a study about classification of snack review and the performance in their research is 80.5% for the average accuracy score. Lexicon-based sentiment analysis method that can be used is Linguistic Inquiry Word Count or LIWC for short. Using the LIWC method, researcher analyzed the sentiment to determine the sentiment of the tweet which can be positive, neutral, or negative sentiment. To find out how the performance of the sentiment analysis is performed, the scores of precision, recall, and f-measure are used as the performance values of the analysis. Research of LIWC have been carried out by [2]. The research conducted a study about comparing text classification method where LIWC is one of them. LIWC method in their research have performed and getting accuracy about 43% to 62.5%. The other research about LIWC have carried out by [1]. The research conducted a study