Analisis Sentimen Pemilihan Calon Presiden Indonesia di Media Sosial Menggunakan Algoritma Support Vector

Abstract

The presidential election held every five years in Indonesia is a crucial component of the democratic process. This study focuses on the significant role of social media, particularly Twitter, as a global expression platform. Utilizing the Support Vector Machine (SVM) algorithm, the research analyzes sentiments related to the Indonesian presidential election through Twitter data. The research methodology involves data collection via Twitter crawling, data preprocessing, data partitioning, as well as the creation, training, and evaluation of the SVM model. The sentiment analysis results indicate a 97% accuracy, with

99% being negative comments. Testing on the dataset of presidential candidate Ganjar Pranowo shows SVM achieving an average accuracy of 98.61%, precision of 98.81%, and recall of 99.79%. This study underscores the effectiveness of SVM in sentiment classification on Twitter data related to presidential elections, offering valuable insights into public perspectives. The findings can be instrumental in gaining a deeper understanding of political dynamics and public opinion.