SAPUTRA, WAHYU BIKA (2026) KLASTERISASI EMISI GAS ATMOSFER AKIBAT KEBAKARAN HUTAN DI KALIMANTAN MENGGUNAKANDATA SENTINEL-5P. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (579kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (153kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (342kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (507kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (645kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (113kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (211kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (381kB) |
Abstract
Forest and land fires in Kalimantan can generate atmospheric gas emissions that may affect environmental conditions. This study aims to cluster areas based on the characteristics of atmospheric gas emissions and environmental factors associated with forest and land fire activities using the K-Means and HDBSCAN algorithms. The main data consist of NO₂, CO, and SO₂ emissions obtained from Sentinel-5P and processed using Google Earth Engine, while supporting data consisting of hotspots, land cover, rainfall, and wind speed were obtained from VIIRS, ESA WorldCover, CHIRPS, and ERA5- Land. The research dataset consists of 52 districts and cities in Kalimantan. The data processing stages include Exploratory Data Analysis (EDA), logarithmic transformation, standardization, and Principal Component Analysis (PCA). Clustering was performed using K-Means and HDBSCAN, while the clustering results were evaluated using the Silhouette Score, Davies-Bouldin Index (DBI), and Calinski-Harabasz Index (CHI). Based on the Elbow Method, K-Means was configured with three clusters, while HDBSCAN produced nine clusters and 17 data points classified as noise. The evaluation results show that HDBSCAN achieved better clustering performance, with a Silhouette Score of 0.385645, a DBI of 0.756857, and a CHI of 19.956910, compared with K-Means, which obtained a Silhouette Score of 0.307256, a DBI of 1.208669, and a CHI of 19.015555. The clustering results were visualized using cluster maps, heatmaps of cluster characteristics, and cluster membership distributions to illustrate the spatial distribution and differences in regional characteristics. The results indicate that HDBSCAN provides more detailed clustering and is capable of identifying areas with distinct characteristics as noise. Keywords: HDBSCAN, K-MEANS, Atmospheric Gas Emissions,Forest Fire,Sentinel-5P, Clustering Kebakaran hutan dan lahan di wilayah Kalimantan dapat menghasilkan emisi gas atmosfer yang berpotensi memengaruhi kualitas lingkungan. Penelitian ini bertujuan untuk mengelompokkan wilayah berdasarkan karakteristik emisi gas atmosfer serta faktor lingkungan yang berkaitan dengan aktivitas kebakaran hutan dan lahan menggunakan algoritma K-Means dan HDBSCAN. Data utama berupa emisi gas NO₂, CO, dan SO₂ diperoleh dari Sentinel-5P dan diolah menggunakan Google Earth Engine, sedangkan data pendukung berupa hotspot, tutupan lahan, curah hujan, dan kecepatan angin diperoleh dari VIIRS, ESA WorldCover, CHIRPS, dan ERA5-Land. Dataset penelitian terdiri atas 52 wilayah kabupaten/kota di Kalimantan. Tahapan pengolahan data meliputi Exploratory Data Analysis (EDA), transformasi logaritmik, standardisasi, dan Principal Component Analysis (PCA). Proses klasterisasi dilakukan menggunakan K-Means dan HDBSCAN, sedangkan evaluasi hasil klasterisasi menggunakan Silhouette Score, Davies-Bouldin Index (DBI), dan Calinski-Harabasz Index (CHI). Berdasarkan Elbow Method, K�Means menggunakan tiga cluster, sedangkan HDBSCAN menghasilkan sembilan cluster dan 17 data yang dikategorikan sebagai noise. Hasil evaluasi menunjukkan bahwa HDBSCAN menghasilkan kinerja klasterisasi yang lebih baik dengan Silhouette Score sebesar 0,385645, DBI sebesar 0,756857, dan CHI sebesar 19,956910, dibandingkan K-Means dengan Silhouette Score sebesar 0,307256, DBI sebesar 1,208669, dan CHI sebesar 19,015555. Hasil klasterisasi divisualisasikan menggunakan peta klaster, heatmap karakteristik klaster, dan distribusi jumlah anggota klaster untuk memperlihatkan persebaran serta perbedaan karakteristik wilayah. Hasil penelitian menunjukkan bahwa HDBSCAN mampu memberikan pengelompokan yang lebih rinci serta mengidentifikasi wilayah yang memiliki karakteristik berbeda sebagai noise. Kata kunci: HDBSCAN, K-MEANS, Emisi Gas Atmosfer,Kebakaran Hutan, Sentinel-5P, Clustering.
Actions (login required)
 |
View Item |
