Sentiment Analysis Based on Appraisal Theory for Assessing Incumbent Electability

Canrakerta Canrakerta (Universitas Indonesia)
Pamuji Lasiyanto Putro (University of Indonesia)
Zikri Irfandi (Universitas Indonesia)
Nur Fitriah Ayuning Budi (University of Indonesia)
Achmad Nizar Hidayanto (University of Indonesia)

Proceeding of the Electrical Engineering Computer Science and Informatics
Vol 5: EECSI 2018

Sentiment analysis is a useful study for determining opinions by classifying text. The document used in the research comes from Twitter about public opinion about community satisfaction related to performance of incumbent. The method used is Appraisal Theory. The data used are 1587 for Jokowi related data, 1774 for Ministry related data, and 1337 government related data. The result of data analysis from this research is that people have positive sentiments for incumbent. Innovation is a term that has the highest positive sentiment, whereas imaging is the term that has the lowest negative sentiment.