Sentiment Analysis for Abolition of National Exams in Indonesia using Support Vector Machine

Through the Minister of Education and Culture, the Indonesian government has announced the abolition of the implementation of the National Examination for grades 6, 9, and 12 in Elementary, Junior, and Senior High Schools. The abolition of this national examination received various responses and comments from numerous groups ranging from leaders to the general public. Therefore, this study aims to apply sentiment analysis in the data mining approach to analyze the textual data of Twitter using a Support Vector Machine (SVM), explore the public’s opinion toward the abolition of the national exam, and measure the policy’s level of acceptance. Furthermore, this study conducted experiments using two other classifiers, namely Random Forest (RF) and Logistic Regression (LR) to deeply observe the performance of SVM. It also conducted scenarios using two different feature extraction, TF-IDF and Bag of Words, and analyzed how it impacts improved accuracy. The experimental results showed that the combination of SVM with Polynomial Kernel and TF�IDF provides performance compared to RF and LR at C=0.01 and degree=20 with accuracy, precision, recall, and F1 score values of 96.97%, 97.28%, 96.87%, and 96.90%, respectively. Furthermore, the result showed that the SVM model with a polynomial kernel provides higher algorithm performance on text classifiers. Therefore, the government can utilize this sentiment analysis as an evaluation material for the decision to abolish the national exam, with most public agreeing with this policy. Index Terms—Data mining, polynomial kernel, sentiment analysis, support vector machine, text classificatio