A Comparative Sentiment Analysis of Public Opinion on Indonesia’s National Football Coach Using CRNN and SVM

This study presents a comparative analysis between a Deep Learning approach using Convolutional Recurrent Neural Network (CRNN) and a Machine Learning approach using Support Vector Machine (SVM) for sentiment classification of public opinion Toward Indonesia’s Football Team New Coach in January 2025 – May 2025 period. The dataset consists of more than 5,100 comments collected from Twitter, preprocessed through normalization, tokenization, stemming, and rebalanced using Synthetic Minority Over-sampling Technique (SMOTE). CRNN was implemented using FastText embeddings and bidirectional GRU layers, while SVM used TF-IDF feature extraction and hyperparameter tuning. Evaluation metrics included accuracy, precision, recall, and F1-score. Experimental results demonstrate that the CRNN model achieved higher performance, particularly in handling informal and context-rich texts, with better generalization in imbalanced data conditions. In contrast, SVM showed competitive performance on structured inputs but struggled in detecting nuanced sentiments. The findings highlight the strength of deep learning models for sentiment analysis in Bahasa Indonesia within social media environments.