Pengelompokan Citra Sumber Karbohidrat Tradisional Kabupaten Belu Berbasis Machine Learning

LEITE, Milu Octaviana Martins Gouveia (2024) Pengelompokan Citra Sumber Karbohidrat Tradisional Kabupaten Belu Berbasis Machine Learning. Undergraduate thesis, Universitas Katolik Widya Mandira Kupang.

	Text ABSTRAK.pdf Download (738kB)
	Text BAB I.pdf Download (287kB)
	Text BAB II.pdf Restricted to Repository staff only Download (263kB)
	Text BAB III.pdf Restricted to Repository staff only Download (479kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (484kB)
	Text BAB V.pdf Download (8kB)
	Text DAFTAR PUSTAKA.pdf Download (411kB)

Abstract

As a regency with considerable agricultural and plantation potential, Belu Regency in East Nusa Tenggara produces various processed carbohydrate sources. Digital identification to support data for the government is necessary. The problem is the lack of identification in the form of machine learning models for images/photos of processed carbohydrate sources that can be used by information technology researchers. The goal is to classify images of processed carbohydrate sources made by the people of Belu Regency using the K-Nearest Neighbor (KNN) method with the help of machine learning technology. The KNN approach is a machine learning algorithm used for classification based on the nearest neighbors of test data. The processed carbohydrate sources to be classified will be separated into five categories: aka bilan, ai uhik hoban, batar sokur, ut fulin, and ut moru. By using feature extraction on a total of 250 images of traditional processed carbohydrate sources, where 70% are training data and 30% are test data, resulting in 175 training images and 75 test images. This study used number of folds 2, 3, 5, and 10, with the highest accuracy achieved at number of fold 10 with an accuracy value of 0.973.

Actions (login required)

View Item