Classification of Pringgasela Typical Songket Using Multi Texton Co-occurrence Descriptor and K-Nearest Neighbor

Abstract

Songket is one of Indonesia's cultural heritages in traditional fabrics that are still preserved today. Pringgasela, a village located on Lombok Island has been producing Songket with distinct characteristics and various patterns. Generally, people are aware of the typical Pringgasela Songket pattern but the difference between one pattern and another is often unrecognized. Furthermore, information regarding the types of Pringgasela Songket has not been well documented. This study aims to build a model that can classify the Pringgalsela Songket patterns using Multi Texton Co-Occurrence Descriptor (MTCD) and K-Nearest Neighbor (KNN) methods. The data used in this study were 4700 images of Pringgasela's Songket, which were divided into training and test data. The highest accuracy obtained was 99.99, 100% precision, and 100% recall with k = 3, using manhattan distance calculation.

Pendeteksian Kecurangan Ujian Melalui CCTV Menggunakan Algoritma YOLOv5

“…Confusion Matrix merupakan salah satu skema evaluasi yang paling sering digunakan dalam machine learning, khususnya supervised learning [28]. Confusion matrix terdiri dari 4 nilai evaluasi, yakni True Positive (TP), True Negatif (TN), False Positif (FP) dan False Negatif (FN) [29].…”