Sentiment Analysis Classification Of Political Parties On Twitter Using Gated Recurrent Unit Algorithm And Natural Language Processing

Abstract

General elections cannot be separated from the issue of political parties. The issue can be in the form of surveys to sentiment. The results of the current survey need to be done in-depth validation related to the truth. Sentiment analysis aims to validate the truth of the survey institution. There are 5 political parties used as datasets in this study, namely Partai Demokrasi Indonesia Perjuangan Party (PDIP), Gerakan Indonesia Raya Party (Gerindra), Golongan Karya Party (Golkar), Partai Kebangkitan Bangsa Party (PKB), and Nasional Demokrat Party (Nasdem). The Gated Recurrent Unit (GRU) algorithm is implemented in this research as an experiment in data calculation. Based on the results of the GRU algorithm calculation in calculating sentiment on political parties, it produces the highest data at 56.50% accuracy, 72.76% precision, and 100% recall