Data-Driven Sentiment Analysis of Public Opinion on National Football Team Coaching Using Naïve Bayes

Abstrak

The appointment of a national football team coach often generates intense public discourse, particularly on social media, where opinions are expressed rapidly and at scale. However, such evaluations are typically subjective and lack a systematic, data-driven foundation. This study proposes a quantitative approach to assess public sentiment toward the Indonesian national team coach using sentiment analysis based on the Naïve Bayes classifier within the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework. A dataset of 2,000 Indonesian-language tweets was collected from social media X and processed through data cleansing, normalization, tokenization, and sentiment labeling. The Naïve Bayes model was applied to classify sentiment polarity, and its performance was evaluated using accuracy, precision, recall, and ROC–AUC metrics. The results show that the model achieves an accuracy of 94.17%, with precision of 92.86% and recall of 94.55%, indicating strong classification performance. However, the AUC value of 0.686 suggests moderate discriminative ability. The findings reveal that public sentiment toward the coach is predominantly positive, reflecting increased supporter confidence. This study demonstrates the effectiveness of data-driven approaches for objective evaluation in sports analytics and highlights opportunities for improvement using advanced machine learning techniques.