PENGEMBANGAN SISTEM PENDUKUNG KEPUTUSAN PEMILIHAN BIBIT UNGGUL SAPI BALI MENGGUNAKAN METODE K-NEAREST NEIGHBOR

Abstract

Bali cattle is an Indonesian native cattle that have a characteristic of the color of his skin. Bali cattle skin color can indicate the quality of the Bali cattle. The classification of the quality of Bali cattle directly is difficult because the human eye has a limited ability to see colors. A decision support system that is able to classify the quality of Bali cattle is based on a digital image of the skin color can help to overcome these limitations. The system will classify Bali cattle into three classes, namely Good (Seeds Superior), Average and Poor. System applying the K-Nearest Neighbor algorithm for the classification process is based on the average features and standard deviation of the red, green, and blue (RGB). This research tested a method to obtain the best value of K, the best image size, and the amount of training data best that will be used. Male Bali cattle using a value of K = 3, image size = 128×128 pixel, and the amount of training data = 45. While the female Bali cattle using a value of K = 6, image size = 64×64 pixel, and the amount of training data = 30. The results of testing the accuracy of the system for male Bali cattle is 100%, while the results of testing the accuracy of the system for female Bali cattle is 66.67%.