Support Vector Machine Algorithms for Sentiment Analysis on the Inaugural Ceremony in the Capital City of Nusantara

Abstract

The inauguration preview at IKN uploaded to the official Instagram account of President Joko Widodo (@jokowi) has garnered various reactions from the public, as seen in the comment columns that expressed all kinds of opinions, from positive to neutral and even negative. The objective of this paper is to provide an objective, systematic and real-time sentiment analysis of public using machine learning algorithms such as Naïve Bayes, and Support Vector Machine (SVM). The analysis process was done on two methods of labelling the data which are VADER and IndoBERT, as well as by implementing both on balancing a class imbalance using SMOTE. The performance result of SVM (RBF kernel) and IndoBERT labelling (95.05% accuracy, 0.951 F1-score) is the best one from evaluation process. The detect result can be analysed in terms of the frequency, which is positive posts and negative comments more than dominating. This work substantiates that such machine learning method, SVM for instance, can be employed to map and interpret the public opinion on government policy based on social media comment data