Classification of Tuberculosis in Rumah Sakit Paru Sumatera Barat Using the C5.0 Algorithm

Abstract

Tuberculosis (TB) remains a serious public health problem, including in West Sumatra Province, where the number of reported cases has continued to increase in recent years. Consequently, effective methods are required to support early detection and accurate classification of TB patients. This study aims to classify the tuberculosis status of patients at Rumah Sakit Paru Sumatera Barat by applying the C5.0 algorithm. The data used in this study consists of secondary data extracted from patient medical records collected from october to december 2024 with a total of 150 patient medical records. The dataset included eight predictor variables representing clinical symptoms and one target variable, namely sputum smear (BTA) examination results. The research process involved data preprocessing, after which the dataset was divided into training and testing subsets using a 70:30 ratio, a classification model was developed using the C5.0 algorithm, and its performance was evaluated using a confusion matrix. The findings indicate that the C5.0 algorithm achieved an accuracy of 91.11%, with a precision of 95.83%, sensitivity of 88.46%, and specificity of 94.74%. Night sweats were identified as the most influential variable in the construction of the decision tree. These findings indicate that the C5.0 algorithm demonstrates excellent performance and can be applied as a decision support method for classifying tuberculosis based on patients’ clinical symptoms