Pemantauan Perubahan Tutupan Lahan Tahun 2019-2024 Menggunakan Algoritma Random Forest dan Extreme Gradient Boosting (Studi Kasus: Pulau Sumbawa)

Raharjo, Aldhian Ramdhany (2025) Pemantauan Perubahan Tutupan Lahan Tahun 2019-2024 Menggunakan Algoritma Random Forest dan Extreme Gradient Boosting (Studi Kasus: Pulau Sumbawa). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5016211092-Undergraduate Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Pulau Sumbawa di Provinsi Nusa Tenggara Barat menghadapi permasalahan serius terkait perubahan tutupan lahan yang dipicu oleh aktivitas pertambangan dan ekspansi pertanian. Data menunjukkan bahwa sekitar 60% hutan di NTB mengalami kerusakan, dengan faktor utama berasal dari sektor pertambangan serta alih fungsi lahan menjadi ladang jagung yang telah mencapai 190.000 hektar hingga tahun 2024. Penelitian ini bertujuan untuk mengklasifikasikan tutupan lahan di Pulau Sumbawa selama periode 2019–2024 menggunakan citra satelit Sentinel-2 dan algoritma Machine Learning, yaitu Random Forest dan Extreme Gradient Boosting (XGBoost). Sentinel-2 dipilih karena keunggulannya dalam resolusi spektral, spasial, dan temporal yang mendukung pemantauan multitemporal. Hasil penelitian menunjukkan bahwa XGBoost menghasilkan akurasi klasifikasi yang lebih tinggi dibandingkan Random Forest, dengan nilai Overall Accuracy sebesar 95,04% dan Kappa 93,28%, sementara Random Forest memperoleh OA 94,63% dan Kappa 92,48%. Pada Algoritma Random Forest, kelas Hutan mengalami penurunan 6,23%, sedangkan pada XGBoost hanya 2,79%. Untuk kelas Vegetasi Non-Hutan, Random Forest mencatat peningkatan sebesar 79,83%, sedangkan XGBoost mengalami peningkatan sebesar 51,99%. Pada kelas Pertanian, Random Forest menunjukkan penurunan sebesar 26,51%, sementara XGBoost turun sebesar 19,24%. Kelas Lahan Terbangun hasil algoritma Random Forest mengalami peningkatan sebesar 5,78%, sedangkan pada XGBoost terjadi penurunan sebesar 20,81%. Untuk Badan Air, Random Forest mengalami penurunan sebesar 43,31%, dan XGBoost turun sebesar hektar 45,83%. Lahan terbuka mengalami peningkatan pada Random Forest sebesar 2.145,73 hektar 39,35%, sementara XGBoost naik sebesar 91,08%.
======================================================================================================================================
Sumbawa Island in West Nusa Tenggara Province is facing serious land cover change issues driven by mining activities and agricultural expansion. Data show that approximately 60% of forests in the province have been degraded, with the main contributing factors is the mining sector and land conversion into cornfields, which has reached 190,000 hectares by 2024. This study aims to classify land cover on Sumbawa Island for the period 2019–2024 using Sentinel-2 satellite imagery and Machine Learning algorithms, namely Random Forest and Extreme Gradient Boosting (XGBoost). Sentinel-2 was chosen due to its advantages in spectral, spatial, and temporal resolution, which support multitemporal monitoring. The results show that XGBoost achieved higher classification accuracy than Random Forest, with an Overall Accuracy of 95.04% and a Kappa coefficient of 93.28%, while Random Forest achieved an OA of 94.63% and a Kappa of 92.48%. In the Random Forest algorithm, the Forest class decreased by 6.23%, whereas in XGBoost it decreased by only 2.79%. For the Non-Forest Vegetation class, Random Forest recorded an increase of 79.83%, while XGBoost showed an increase of 51.99%. The Agricultural class decreased by 26.51% in Random Forest and 19.24% in XGBoost. The Built-up Land class increased by 5.78% in Random Forest, but decreased by 20.81% in XGBoost. Water Bodies decreased by 43.31% in Random Forest and 45.83% in XGBoost. Lastly, Open Land increased by 2,145.73 hectares (39.35%) in Random Forest, and by 91.08% in XGBoost.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Random Forest, Extreme Gradient Boosting, Sentinel-2, Tutupan Lahan, Random Forest, Extreme Gradient Boosting, Sentinel-2, Land Cover.
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.217 Geospatial data G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing
Divisions:	Faculty of Civil Engineering and Planning > Geomatics Engineering > 29202-(S1) Undergraduate Thesis
Depositing User:	Aldhian Ramdhany Raharjo
Date Deposited:	24 Jul 2025 07:11
Last Modified:	24 Jul 2025 07:11
URI:	http://repository.its.ac.id/id/eprint/121343

Actions (login required)

View Item