Sentiment Analysis of Online Learning Students Feedback for Facing New Semester: A Support Vector Machine Approach

Abstract

Students often experience various feelings when facing a new semester. Feelings such as anxiety, fear, and excitement can occur when students take classes in online learning. As a result, in the new semester period, students gave various responses to their online lectures in the new semester. This study aims to classify online class student feedback on their participation in the new semester. Questionnaires in the form of essay questions were distributed to 375 students who took online lectures in the 2nd semester of the 2020/2021 academic year to find out how they felt about attending online lectures in the new semester. This study uses the sentiment analysis method to identify and extract student responses in subjective information that focuses on positive and negative polarities. The findings of this study indicate that sentiment analysis using the Support Vector Machine (SVM) method produces an accuracy of 84%. SVM produces a positive prediction precision value of 77.61%, while the negative predictive precision value gets 94.26%. The experimental results show that sentiment analysis using the SVM method can classify student responses based on two polarities, namely positive and negative polarity.