<![CDATA[This study aims to analyze public sentiment towards the Indonesian National Team during the 2023 Asian Cup through Instagram comments, using the Indonesian-language Transformer model. A total of 21,045 comments were collected from the official @timnasindonesia account and filtered into 17,829 comments worthy of analysis after going through preprocessing processes such as text cleaning, case folding, normalization, tokenization, stopword removal, and stemming. Comments were then automatically classified into three sentiment classes, namely positive, negative, and neutral, using the Indonesian Sentiment Lexicon (InSet). Three Transformer models were used, namely IndoBERT, IndoRoBERTa, and DistilBERT Multilingual, and compared to the baseline SVM + TF-IDF. The evaluation used accuracy, precision, recall, and F1-score metrics. The results showed that IndoBERT (learning rate 5e-5) gave the best performance with an accuracy of 0.8897 and an F1-score of 0.8859, outperforming other models. The analysis was conducted by considering the typical challenges of Instagram comments such as slang, abbreviations, emojis, and mixed use of Indonesian-English. These findings validate the effectiveness of the monolingual Transformer model for Indonesian, which is still rarely compared systematically in the context of social media. These findings can also be used by the management of the Indonesian National Team or sports policy makers to evaluate public responses in real-time, as well as being a reference in developing an adaptive and contextual public opinion analysis system.]]>
Copyrights © 2025
