Sentiment Analysis for Online Learning using The Lexicon-Based Method and The Support Vector Machine Algorithm

Abstract

The pros and cons regarding online learning has been a hot topic in society, both on social media and in the real world. Indonesian netizens still post opinions about online learning on social media such as Twitter. This study aims to analyze public comments to determine whether the trend of the comments is positive, negative, or neutral. The classification of netizen opinions is called sentiment analysis. This study applies 2 ways of carrying out sentiment analysis. The first stage employs the SVM algorithm with data labeling automatically obtained from the Emprit Academy drone portal while the second stage is still using the SVM algorithm but the data labeling with lexicon-based method. The results of this study are comparisons of labels obtained automatically from the Emprit Academy drone portal and labeling using lexicon based. The SVM algorithm obtains an accuracy of 90%, while the use of lexicon-based increases the accuracy value by 5% to 95%. It can be concluded that labeling data using a lexicon-based method can improve the accuracy of the SVM algorithm.

Development and Comparison of Multiple Emotion Classification Models in Indonesia Text Using Machine Learning

“…Furthermore, while TF-IDF produced suboptimal accuracy results, each algorithm can be enhanced through various methods. In particular, this study should address data imbalance before accurately assessing performance, with data balancing achievable through techniques such as SMOTE [7].…”

“…Manual classification is usually conducted by humans without the help of intelligent computer algorithms while ethnology-assisted classification uses several algorithms such as Naï ve Bayes Algorithm, Support Vector Machine, Decision Tree, Fuzzy Logic, and Artificial Neural Networks [5,6]. Previous research has also discussed various data for classification using the Indonesian language [7].…”