Comparison of Support Vector Machine (SVM) and Random Forest Classifier Algorithms in the Classification of Acute Respiratory Infection (ARI) Disease

The increasing number of Acute Respiratory Infection (ARI) cases in Jabodetabek, averaging 200 thousand per month, is linked to rising air pollution levels in Jakarta. This study compares classification models using Support Vector Machine (SVM) and Random Forest algorithms for classifying ARI, specifically nasopharyngitis and pharyngitis, based on data from the Matraman District Health Center (2021–2023). The models employed both default settings and hyperparameter tuning with resampling techniques such as undersampling and oversampling to address class imbalance. Experimental results show that SVM achieved 98% accuracy with a sample size of 500, while Random Forest reached 96.4% accuracy with a sample of 2500. Hyperparameter tuning had minimal impact on improving accuracy. Despite resampling efforts, the precision for pharyngitis remained lower. This research highlights the efficacy of SVM and Random Forest for ARI classification, providing a basis for further investigation and practical application in improving diagnostic accuracy for respiratory infections.