Analisis Sentimen Terhadap Program Indonesian International Student Mobility Award (Iisma) Di Platform X / Twitter Menggunakan Tf-Idf Dan Svm

Abstract

Indonesian International Student Mobility Award (IISMA) is a Merdeka Belajar Kampus Merdeka (MBKM) program aimed at broadening perspectives and fostering international relations for Indonesian students. This study analyzes public sentiment towards IISMA on platform X using sentiment analysis based on Support Vector Machine (SVM) with linear, polynomial, and Radial Basis Function (RBF) kernels. The dataset was obtained through web scraping and processed using text mining techniques with TF-IDF feature extraction. The results indicate that the SVM model can classify sentiment into three classes: positive, neutral, and negative, although it struggles to predict the negative class due to dataset imbalance. The RBF kernel achieved the best kinerjance with a precision of 70,52% and an accuracy of 62,92%, while the linear kernel produced the highest F1-score of 53%, indicating a better balance between precision and recall. Performance testing on Indonesian and English text showed better results for English text, with an accuracy of 71,11% compared to 62,44% for Indonesian text, supported by TextBlob’s ability to better recognize sentiment patterns in English. These findings highlight the importance of tailored strategies to improve sentiment analysis kinerjance for Indonesian text, such as manual labeling or tools that support the Indonesian language.