Fake News Detection Using Sentiment Analysis Approach in Indonesian Language

In today’s digital age, information spreads rapidly. It makes news and any other information sources easily accessible and shareable. However, this convenience also comes with the risk of encountering false as well as fake information. This research aims to develop a system for detecting fake news using sentiment analysis approach. While earlier studies successfully used the VADER (Valence Aware Dictionary and Sentiment Reasoner) method for English and achieving a noteworthy 78% accuracy with Random Forest, our focus is on designing a fake news detection system based on sentiment analysis for the Indonesian language. We extract extreme sentiment scores from news headlines and employ Random Forest, Support Vector Machine, and Naive Bayes for classification. Results from our experiment show Random Forest yields promising results, averaging 62% accuracy and peaking at 66%. This is a positive step in Indonesian fake news identification through sentiment analysis. Moving ahead, improving this system could greatly combat the spread of fake news. Future efforts may refine sentiment analysis, explore advanced ML techniques, and adapt to emerging news trends. Developing a strong fake news detection system via sentiment analysis contributes to promoting accurate information in our connected world.