Comparative Analysis of Support Vector Machine and Random Forest Algorithms in Indonesian Batik Classification

Abstract

This study compares the performance of Support Vector Machine (SVM) and Random Forest (RF) algorithms in Indonesian batik image classification. Data collected from four batik categories: Pattern Batik Insang, Pattern Batik, Patterns Batik Dump, and Pattern Megamendung. Image feature extracted using Histogram of Oriented Gradients (HOG). SVM models with linear and RF kernels with 100 decision trees are trained and tested on this dataset. The evaluation results showed that the SVM has an accuracy of 88%, with precision and recall balanced between classes, while RF has an accuracy of 86%, with some classes showing excellent performance. SVM is superior in overall accuracy, but RF offers better interpretability and ease of setting parameters. The conclusions of this study suggest that both algorithms are able to effectively classify bacterial images, but the selection of the algority depends on the specific needs of the application. Further adjustment of parameters and additional preprocessing techniques are recommended to improve model performance. This research provides a strong foundation for further development in the classification of batic images using machine learning.