<![CDATA[The rapid growth of the esports industry in Indonesia presents unique challenges for professional teams such as EVOS Esports, particularly in strengthening fan engagement and loyalty in the digital era. This study aims to analyze fan sentiment toward the official Instagram posts of EVOS Esports using a deep learning approach with a Convolutional Neural Network (CNN). The research process involved data collection through web scraping, followed by preprocessing stages such as cleaning, case transformation, normalization, tokenization, stopword removal, and stemming. The dataset was then labeled, split into training and testing sets (90:10), and used for CNN model training and evaluation through a confusion matrix. The results demonstrate that the CNN model successfully classified comments into three sentiment categories—positive, negative, and neutral—with an accuracy of 92%. The model also achieved a precision of 0.92, recall of 0.92, and an F1-score of 0.92, indicating very good classification performance. Sentiment distribution analysis of 11,305 comments showed that neutral sentiment dominated (47.24%), followed by positive (30.12%) and negative (22.64%). These findings provide valuable insights into fan perceptions of esports team performance on social media. For future research, expanding the sentiment lexicon with terms commonly used in online communities is recommended to further enhance classification accuracy.]]>
Copyrights © 2025
