<![CDATA[Twitter is one of the most dynamic social media platforms that provides real-time information through its trending topics feature, which reflects the most talked about issues among users. However, in Indonesia, trending topics are often dominated by entertainment, celebrity gossip or light-hearted viral content, and are not used to highlight or analyze more substantial social issues. This study aims to classify Twitter trending topics in Indonesia using three clustering algorithms: K-Means, DBSCAN, and Latent Dirichlet Allocation (LDA). Data was collected over a certain period and processed through a text preprocessing stage before applying the clustering algorithms. The results show that LDA without keyword filtering provides the most relevant and dominant topic classification, the bar chart results tend to be dominant in topic 0 there are as many as 160 topics with the main cluster relating to the Indonesian presidential election. These findings suggest that LDA outperforms K-Means and DBSCAN in identifying latent topic structures in Twitter data. This study contributes to a better understanding of trending topics and supports data-driven public opinion analysis and decision-making.]]>
Copyrights © 2025
