Implementasi Klasifikasi Jenis Kayu Menggunakan Algoritma KNN dan Gray Level Cooccurence Matrix

Abstract

This study aims to obtain classification accuracy results derived from testing various K parameters. The research method uses the K-Nearest Neighbor algorithm as the determinant of classification results, supported by the Gray Level Co-occurrence Matrix (GLCM) for feature extraction. Through GLCM feature extraction, grayscale values have been obtained and used as data to determine the final nearest neighbor. The testing is limited to using only odd-numbered K parameters, totaling eight K parameters. This study employs K parameters: K=1, K=3, K=5, K=7, K=9, K=11, K=13, and K=15, using a dataset of 150 wood images. The wood images are divided into three different wood types that have been collected. The dataset is divided into 120 training data and 30 testing data. The testing results show that the system achieves the lowest accuracy of 76.67% when using parameter K=13, while the highest accuracy of 90.00% is achieved when using parameter K=1. The average accuracy across all tests is 83.75%