<![CDATA[Hotel building maintenance in tourism-driven regions such as the Special Region of Yogyakarta is essential for maintaining service quality and building reliability. However, conventional assessment methods often depend on subjective expert judgment, which may be time-consuming and inconsistent. This study evaluates the performance of Support Vector Machine (SVM) combined with Principal Component Analysis (PCA) to classify hotel building maintenance levels in Yogyakarta into five categories. The dataset includes 175 samples and 11 variables covering architectural, structural, mechanical, electrical, outdoor space, and housekeeping aspects, with naturally imbalanced class distribution. Data were normalized using MinMaxScaler and reduced to nine principal components, explaining 93.41% of cumulative variance. Three SVM kernels polynomial, radial basis function (RBF), and sigmoid were tested using a 70:30 training–testing split with default scikit-learn hyperparameters. The sigmoid kernel produced the best performance, achieving 90.57% accuracy, 91.32% precision, 90.57% recall, and 90.53% F1-score, outperforming RBF and polynomial kernels. Stratified 5-Fold cross-validation showed an average accuracy of 81.71% ± 9.66%. The results indicate that PCA-SVM with a sigmoid kernel is effective for automated hotel building maintenance classification.]]>
Copyrights © 2026
