PUBLIC SENTIMENT ANALYSIS OF THE INDONESIAN NATIONAL FOOTBALL TEAM ON INSTAGRAM USING NAIVE BAYES

Abstract

Social media platforms such as Instagram have become essential spaces for fans of the Indonesian National Football Team to articulate their reactions and support. Systematically analyzing these comments offers a valuable window into the collective sentiment and perception surrounding match outcomes and team management decisions. This research applies the Naive Bayes Classifier (NBC), a widely recognized probabilistic method for text classification, to categorize Instagram comments into sentiment classes. The NBC operates on the principle of conditional probability and is particularly advantageous due to its efficiency with limited training data. Between September 5, 2024, and March 25, 2025, a dataset of 1,500 comments was assembled from 15 Instagram posts reflecting different match results—victories, draws, and defeats. The analysis revealed that while NBC attained an overall classification accuracy of 90%, its performance varied across sentiment categories. The model was especially adept at identifying Neutral comments, achieving high precision and recall, but demonstrated limitations in reliably classifying Positive and Negative sentiments. These findings highlight both the potential and the challenges of deploying NBC for sentiment analysis in imbalanced social media datasets.