<![CDATA[Indonesia is a country prone to natural disasters, especially earthquakes. One of the biggest threats is the potential for megathrust earthquakes that can cause severe damage and loss of life, especially in the Jakarta area. In the digital era, information about the potential for megathrust earthquakes is widely disseminated through social media platforms such as YouTube, which then triggers various opinions and comments from the public. A video titled "MEGATHRUST EARTHQUAKE!! JAKARTA RESIDENTS MUST BE PREPARED FOR THE COLLAPSE OF BUILDINGS?" uploaded by the Kamar Jeri YouTube channel has attracted public attention and sparked discussions about community preparedness for this potential disaster. This study aims to analyze public opinion sentiment towards the video using two machine learning methods, namely Naive Bayes and Support Vector Machine (SVM). To overcome the imbalance of data between sentiment classes, the Synthetic Minority Over-sampling Technique (SMOTE) technique was applied. The results showed that SMOTE was effective in improving the performance of both models, but the improvement in SVM was more significant. SVM performed better than Naive Bayes in classifying public opinion sentiment]]>
Copyrights © 2024
