Pengklasifikasian Citra Dataset Kain Tenun Manggarai Berbasis Machine learning

KURNIATI, Marselina Selviana (2024) Pengklasifikasian Citra Dataset Kain Tenun Manggarai Berbasis Machine learning. Undergraduate thesis, Universitas Katolik Widya Mandira Kupang.

	Text ABSTRAK.pdf Download (623kB)
	Text BAB I.pdf Download (134kB)
	Text BAB II.pdf Restricted to Repository staff only Download (172kB)
	Text BAB III.pdf Restricted to Repository staff only Download (318kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (514kB)
	Text BAB V.pdf Restricted to Repository staff only Download (84kB)
	Text BAB VI.pdf Download (15kB)
	Text DAFTAR PUSTAKA DAN SURAT KETERANGAN HASIL CEK PLAGIASI.pdf Download (276kB)

Abstract

Indonesia is a country rich in cultural heritage, one of which is the diversity of traditional fabrics, especially ikat woven fabrics. However, with the many variations in motifs and colors, it is difficult for the younger generation to distinguish the types of Manggarai woven fabrics based on motifs and colors. This study aims to distinguish the types of Manggarai woven fabrics based on the types of fabric motifs and colors. This study uses 1000 datasets of woven fabric images divided into four classes, namely Songke Cibal, Songke Lambaleda, Songke Ruis, and Songke Todo, and uses the K-Nearest Neighbor and Support Vector Machine methods. The calculation results from the 20-fold cross-validation test with the K-Nearest Neighbor method produced an accuracy of 96.3%, a precision of 96.4%, a recall of 96.3%, and an f1-score of 96.3%, while the Support Vector Machine produced an accuracy of 87.7%, a precision of 89.5%, a recall of 87.7%, and an F1-score of 87.5%. From the results of the tests carried out, the comparison with the highest accuracy, namely the 20-fold cross-validation test on the K-Nearest Neighbor method, produced an accuracy of 96.3%, so that the K-Nearest Neighbor method became the best method in classification to predict Songke Cibal, Songke Lambaleda, Songke Ruis, and Songke Todo fabrics.

Actions (login required)

View Item