Best Algorithm in Sentiment Analysis of Presidential Election in Indonesia on Twitter

The election of presidential candidates for 2024 is included in the democratic process to elect the president and vice president for the period 2024-2029. In this case, there are already names of presidential candidates who have been nominated and many survey institutions have published survey results on several candidates who are eligible to become presidential candidates, based on this, not a few netizens have expressed their opinions that can be made regarding public sentiment. about the trend of presidential candidates which is currently being discussed on Twitter social media. In this study, public sentiment analysis was carried out on trends in presidential candidates by comparing three classification algorithms, namely support vector machine (SVM), K-Nearest Neighbor (K-NN) and Naïve Bayes (NB). Comparisons are made to find out which algorithm has better accuracy. This research is also expected to provide references and knowledge to the public about the trends of presidential candidates in the upcoming presidential election. The data taken are 9966 twitter data regarding presidential and presidential candidates as well as tweet data taken in the second week of 09-17 September 2022. The results of this test concluded that the SVM algorithm is superior to K-NN and Naïve Bayes which get an accuracy rate of 79.57%. The results of this study get the best and most effective algorithm in classifying positive and negative comments on the 2024 presidential candidate trend.

Keywords: Capres, Pilpres, v, SVM, Naive Bayes, KNN