PERBANDINGAN KLASIFIKASI GENUS/SPECIES BUNGA MENGGUNAKAN KOMBINASI GLOBAL FEATURE DESCRIPTION DAN K-NEAREST NEIGHBOUR (K-NN) DAN RANDOM FOREST (RF)

JANNATI, SELFIA and Sukemi, Sukemi and Sutarno, Sutarno (2021) PERBANDINGAN KLASIFIKASI GENUS/SPECIES BUNGA MENGGUNAKAN KOMBINASI GLOBAL FEATURE DESCRIPTION DAN K-NEAREST NEIGHBOUR (K-NN) DAN RANDOM FOREST (RF). Undergraduate thesis, Sriwijaya University.

[thumbnail of RAMA_56201_09011281320030.pdf] Text
RAMA_56201_09011281320030.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (4MB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_TURNITIN.pdf] Text
RAMA_56201_09011281320030_TURNITIN.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (4MB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_01_fornt_ref.pdf]
Preview
Text
RAMA_56201_09011281320030_0003126604_0201117802_01_fornt_ref.pdf - Accepted Version
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Preview
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_02.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_02.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_03.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_03.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_04.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_04.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (792kB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_05.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_05.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (11kB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_06_ref.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_06_ref.pdf - Bibliography
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (80kB) | Request a copy
[thumbnail of RAMA_56201_09011281320030_0003126604_0201117802_07_lamp.pdf] Text
RAMA_56201_09011281320030_0003126604_0201117802_07_lamp.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (236kB) | Request a copy

Abstract

The need for acceleration of the selection or sorting of goods has been and is being developed by industry players both at home and abroad. Products that are selected or sorted are very diverse, one of which is flowers. The selection is based on color, texture and shape because the dataset used has colors that are almost close to more than 1 type of flower so that the combination with shape and texture will make a difference to the type of flower and several studies have shown that the combination of color and texture has been proven successful in finding similar images. Image testing uses primary datasets totaling 80 images per each class, thus there are 55 ~ 54 training data and 25 ~ 26 test data (random images per each class).This study uses 3 methods to take image features with 2 stages that are distinguished from input, namely for RGB images that are converted to grayscale channels, executed by Haralick Texture and Hu Moments, while the complete RGB image is executed by Color Histogram (which in this case RGB to HSV is a characteristic. As a whole, it is taken based on the color for the object (flower) then the classification will be carried out using the k-Nearest Neighbor method and compared with the Random Forest method. Based on the test results it is found that k-Nearest Neighbor with k = 3 produces a higher predictive value of 55% compared to k = 5 and k = 7, which both produce a predictive value of 53% of the 5 flower classes tested 2. Then the comparison of methods is carried out to get a better result increase. Random Forest (RF) produces a better predictive value of 92 % with the highest precision, namely Hyacinthoides L., the highest recall was Tussilago farfara L. (species), and the highest F1-Score was Hyacinthoides L. in 5 flower classes.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Bunga, Color Histogram, Haralick Texture, Hu Moments, Klasifikasi, k-Nearest Neighbour, Random Forest
Subjects: Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions: 09-Faculty of Computer Science > 56201-Computer Systems (S1)
Depositing User: Users 9773 not found.
Date Deposited: 14 Jan 2021 03:47
Last Modified: 14 Jan 2021 03:47
URI: http://repository.unsri.ac.id/id/eprint/39910

Actions (login required)

View Item View Item