<![CDATA[Tea is a commodity with high economic value, but it is susceptible to diseases such as Brown Blight, Red Rust, and Red Spider Mite. The manual identification process currently relies on visual observation, which is time-consuming and prone to error. This research aims to analyze the performance of K-NN and SVM algorithms in classifying tea leaf diseases based on digital images. This research utilized a perfectly balanced dataset of 5,000 images. The research methodology involves image preprocessing and classification using 5-Fold, 10-Fold, and 20-Fold Cross-Validation. The results demonstrate that the SVM algorithm consistently outperforms K-NN across all testing scenarios. Specifically, SVM achieved its highest accuracy of 96.6% using 20-Fold Cross-Validation, whereas the highest accuracy for K-NN was 96.1%. The research concludes that SVM provides superior sensitivity and accuracy for identifying tea leaf diseases, offering a viable solution for automated detection systems in the plantation sector]]>
Copyrights © 2026
