Perbandingan Algoritma Random Forest dan SVM untuk Klasifikasi Tutupan Lahan pada Citra Satelit Sentinel-2 di Kawasan IKN Menggunakan Google Earth Engine

Al Fathoni, Hanif (2025) Perbandingan Algoritma Random Forest dan SVM untuk Klasifikasi Tutupan Lahan pada Citra Satelit Sentinel-2 di Kawasan IKN Menggunakan Google Earth Engine. Undergraduate thesis, UPN Veteran Jawa Timur.

This is the latest version of this item.

	Text (Cover) 20081010001-COVER.pdf Download (1MB)
	Text (Bab 2) 20081010001-BAB 2.pdf Restricted to Repository staff only until 27 November 2027. Download (631kB)
	Text (Bab 3) 20081010001-BAB 3.pdf Restricted to Repository staff only until 27 November 2027. Download (586kB)
	Text (Bab 4) 20081010001-BAB 4.pdf Restricted to Repository staff only until 27 November 2027. Download (4MB)
	Text (Bab 5) 20081010001-BAB 5.pdf Restricted to Repository staff only until 27 November 2027. Download (120kB)
	Text (Daftar Pustaka) 20081010001-DAFTAR PUSTAKA.pdf Download (176kB)
	Text 20081010001-LAMPIRAN.pdf Restricted to Repository staff only Download (7MB)
	Text (Bab 1) 20081010001-BAB 1.pdf Download (115kB)

Abstract

Land cover analysis plays an important role in supporting sustainable regional development planning. This study aims to classify land cover in the Nusantara Capital City (IKN) area using Sentinel-2 imagery processed on the Google Earth Engine (GEE) platform. Two supervised learning algorithms, Random Forest (RF) and Support Vector Machine (SVM), were employed with feature extraction from spectral indices (NDVI, NDBI, NDWI) and cloud masking techniques to reduce atmospheric disturbance. The results show that Random Forest with 50 trees (RF_50trees) achieved the best performance, with a validation accuracy of 88%, an F1-score of 88.03%, the fastest computation time (26.48 seconds), and the highest efficiency (0.0332 accuracy/second). The SVM Linear model with cost=10 (SVM_LINEAR_cost10) also demonstrated competitive results, reaching an accuracy of 86.67% and an F1-score of 86.68%, although requiring longer computation time (57.14 seconds) with lower efficiency. Producer’s Accuracy (PA) and User’s Accuracy (UA) analysis indicated that RF was more stable across all classes, while SVM was more sensitive, particularly in distinguishing open land, which was often misclassified as built-up land. Overall, the study concludes that Random Forest (RF_50trees) is the most optimal algorithm for land cover classification in IKN, while SVM Linear (cost=10) can still serve as a competitive alternative.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Junaidi, Achmad	NIDN0710117803	achmadjunaidi.if@upnjatim.ac.id
Thesis advisor	Aditiawan, Firza Prima	NIDN0434010322	firzaprima.if@upnjatim.ac.id

Subjects:

T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Mr Hanif Al Fathoni

Date Deposited:

27 Nov 2025 08:14

Last Modified:

27 Nov 2025 08:14

URI:

https://repository.upnjatim.ac.id/id/eprint/47103