KOMPARASI ALGORITMA K-NEAREST NEIGHBOR (KNN), SUPPORT VECTOR MACHINE (SVM) DAN NAIVE BAYES UNTUK KLASIFIKASI KEMATANGAN BUAH SAWIT

Safrida Ika Guslianto, S.T, NIM.: 21206051001 (2023) KOMPARASI ALGORITMA K-NEAREST NEIGHBOR (KNN), SUPPORT VECTOR MACHINE (SVM) DAN NAIVE BAYES UNTUK KLASIFIKASI KEMATANGAN BUAH SAWIT. Masters thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (KOMPARASI ALGORITMA K-NEAREST NEIGHBOR (KNN), SUPPORT VECTOR MACHINE (SVM) DAN NAÏVE BAYES UNTUK KLASIFIKASI KEMATANGAN BUAH SAWIT) 21206051001_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf - Published Version Download (3MB) | Preview
	Text (KOMPARASI ALGORITMA K-NEAREST NEIGHBOR (KNN), SUPPORT VECTOR MACHINE (SVM) DAN NAÏVE BAYES UNTUK KLASIFIKASI KEMATANGAN BUAH SAWIT) 21206051001_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf - Published Version Restricted to Registered users only Download (4MB) | Request a copy

Abstract

The development of digital technology is widely applied to the world of agriculture, one of which is artificial intelligence. Computer Vision is part of artificial intelligence which in the learning process uses visual data. The introduction of the maturity level of oil palm fruit is a form of collaboration between technology and the world of agriculture. Several previous studies have discussed the algorithms used in classifying oil palm fruit. Previous research was dominated by the introduction of maturity levels using color features using algorithms such as Backpropogation, K-Nearest Neighbor, Decession Tree and SVM. To obtain the optimal and relevant algorithm for classifying the maturity level of oil palm fruit, it is necessary to compare the classification algorithm and the features used. In this study, the authors are interested in making comparisons of algorithms using color, texture and shape features. The algorithms being compared are the KNN, SVM and Naïve Bayes algorithms. The purpose of this research is to find out the best features and algorithms that get the highest accuracy. The dataset used is image data of unripe palm, ripe palm and rotten palm with a total of 600 training images and 150 test images. This study uses color features, namely RGB average, RGB Standard Deviation, RGB Skewness and RGB Entropy. Texture features are Grayscale Average, Grayscale Standard Deviation, Contrast, Correlation, Energy and Homoginety. The shape features are Area, Perimeter, Metrics, Major axis, Minor axis and Eccentricity. The results of this study obtained the highest accuracy obtained by the KNN Algorithm with k = 3, on color features = 91.3%, texture features = 89.3%, shape features = 83.3% and all features = 98%. The SVM algorithm obtains an accuracy of color features = 83.3%, texture features = 50%, shape features = 47.3% and all features = 42%. Whereas Naïve Bayes on color features = 65%, texture features = 68, shape features = 42 and all features = 77%.

Share this knowledge with your friends :

Actions (login required)

View Item