The Accuracy Value Comparison of Sentiment Analysis on 2024 General Election Keywords

Abstract

Democracy is a government system that gives equal rights to every citizen in making decisions that can affect their lives. In Indonesia, the democratic system is realized through the General Election (Pemilu) organization, which is held periodically. An election is always discussed in the real and virtual worlds, especially on X social media. Considering the huge number of tweets, manual analysis of public opinion is inefficient. Therefore, technology is needed to analyze and categorize tweets into positive or negative sentiments automatically. This research compared the accuracy value of sentiment analysis on 2024 election data keywords using the NAIVE BAYES CLASSIFIERS and SUPPORT VECTOR MACHINE methods. The data used was 5651 tweets and obtained an accuracy value of 64.59% in the naïve bayes classifiers method and 76.14% in the support vector machine method. It shows that SVM is reliable in the context of sentiment analysis involving complex and diverse data.