<![CDATA[The issue of lesbian, gay, bisexual, and transgender in Indonesia remain largely taboo and often provoke various public reactions, particularly on social media platforms. This phenomenon highlights the need to map public opinion through sentiment analysis. This study aims to compare the performance of three Indonesian-language transformer models IndoBERT, IndoRoBERTa, and NusaBERT in analyzing sentiment related to LGBT issues on platform X. The research methodology follows the Cross Industry Standard Process for Data Mining approach. Testing was conducted using 6,000 preprocessed and automatically labeled data points, showing that the IndoBERT model achieved the best performance with an accuracy of 90.50% and an F1-score of 56.88%. This best-performing model was then used to classify 12,804 tweets, revealing that 92.80% expressed negative sentiment, 6.71% neutral, and only 0.49% positive. These findings confirm that negative public perceptions of LGBT issues remain highly dominant in Indonesia’s digital space.]]>
Copyrights © 2025
