ANALISIS DEFORMASI PERMUKAAN PADA BULAN JULI–DESEMBER 2024 DI WILAYAH SEKITAR SEGMEN RAKUTAI MENGGUNAKAN ALGORITMA RANDOM FOREST DAN MULTI-LAYER PERCEPTRON

    Hanipah Nurdini, - and Hendro Murtianto, - and Silmi Afina Aliyan, - (2025) ANALISIS DEFORMASI PERMUKAAN PADA BULAN JULI–DESEMBER 2024 DI WILAYAH SEKITAR SEGMEN RAKUTAI MENGGUNAKAN ALGORITMA RANDOM FOREST DAN MULTI-LAYER PERCEPTRON. S1 thesis, Universitas Pendidikan Indonesia.

    Abstract

    Deformasi permukaan (ground deformation) merupakan ancaman geodinamika serius karena dapat merusak infrastruktur vital, mengganggu aktivitas sosial-ekonomi, dan mengancam keselamatan masyarakat di wilayah jalur sesar aktif. Di Indonesia, ancaman ini semakin relevan karena banyak kawasan padat penduduk berada di atas struktur geologi aktif, termasuk Sesar Garsela di Jawa Barat—khususnya Segmen Rakutai yang memicu dua gempa pada 18 September dan 7 Desember 2024. Penelitian ini bertujuan untuk mengkelaskan tingkat deformasi permukaan menggunakan algoritma Random Forest dan MLP berdasarkan nilai LOS displacement hasil DInSAR, serta menganalisis pola spasial deformasi yang terbentuk pada wilayah studi. Hasil pemodelan deformasi dengan algoritma Random Forest (RF) dan Multi-Layer Perceptron (MLP) berhasil memetakan zonasi deformasi permukaan ke dalam lima kelas dengan pola spasial yang jelas. Kelas deformasi sedang mendominasi wilayah studi, masing-masing seluas 37.002 ha (RF) dan 38.464 ha (MLP), terutama tersebar di area yang berdekatan dengan Segmen Rakutai. Pada sisi lain, kelas deformasi tinggi teridentifikasi secara terbatas namun lebih terlokalisasi di zona yang berdekatan dengan struktur sesar dan area berlereng curam, dengan luas mencapai 5.418 ha (RF) dan 4.436 ha (MLP). Sementara itu, MLP menunjukkan kecenderungan lebih responsif dalam memetakan uplift tinggi (2.300 ha) dibandingkan RF (1.795 ha). Temuan ini menunjukkan bahwa subsiden lebih mendominasi luasan wilayah, sedangkan uplift muncul secara spasial sebagai pola-pola kecil yang terfragmentasi. Kata kunci: deformasi permukaan, DInSAR, Sentinel-1, Random Forest, Multi-Layer Perceptron, Sesar Garsela Ground deformation is a serious geodynamic threat because it can damage vital infrastructure, disrupt socio-economic activities, and threaten the safety of people living in active fault zones. In Indonesia, this threat is increasingly relevant because many densely populated areas are located above active geological structures, including the Garsela Fault in West Java—particularly the Rakutai Segment, which triggered two significant earthquakes in September 18 and December 7, 2024. This study aims to classify the level of surface deformation using the Random Forest and MLP algorithms based on the LOS displacement values resulting from DInSAR, and analyze the spatial patterns of deformation formed in the study area. The results of deformation modeling using the Random Forest (RF) and Multi-Layer Perceptron (MLP) algorithms successfully mapped the surface deformation zoning into five classes with clear spatial patterns. The moderate deformation class dominates the study area, covering 37,002 ha (RF) and 38,464 ha (MLP), respectively, mainly spread in areas adjacent to the Rakutai Segment. On the other hand, high deformation classes were identified in a limited but more localized manner in zones adjacent to fault structures and steep-slope areas, with an area reaching 5,418 ha (RF) and 4,436 ha (MLP). Meanwhile, the MLP showed a more responsive tendency in mapping high uplift (2,300 ha) compared to the RF (1,795 ha). This finding indicates that subsidence dominates the area, while uplift appears spatially as small, fragmented patterns. Keywords: surface deformation, DInSAR, Sentinel-1, Random Forest, Multi-Layer Perceptron, Garsela Fault

    Baca Full Text klik disini
    [thumbnail of S_SIG_2104812_Title.pdf] Text
    S_SIG_2104812_Title.pdf
    Download (775kB)
    [thumbnail of S_SIG_2104812_Chapter1.pdf] Text
    S_SIG_2104812_Chapter1.pdf
    Download (583kB)
    [thumbnail of S_SIG_2104812_Chapter2.pdf] Text
    S_SIG_2104812_Chapter2.pdf
    Restricted to Staf Perpustakaan
    Download (870kB)
    [thumbnail of S_SIG_2104812_Chapter3.pdf] Text
    S_SIG_2104812_Chapter3.pdf
    Download (4MB)
    [thumbnail of S_SIG_2104812_Chapter4.pdf] Text
    S_SIG_2104812_Chapter4.pdf
    Restricted to Staf Perpustakaan
    Download (5MB)
    [thumbnail of S_SIG_2104812_Chapter5.pdf] Text
    S_SIG_2104812_Chapter5.pdf
    Download (114kB)
    [thumbnail of S_SIG_2104812_Appendix.pdf] Text
    S_SIG_2104812_Appendix.pdf
    Restricted to Staf Perpustakaan
    Download (879kB)
    Official URL: https://repository.upi.edu/
    Item Type: Thesis (S1)
    Additional Information: https://scholar.google.com/citations?hl=id&user=RwD7HI0AAAAJ ID SINTA Dosen Pembimbing Hendro Murtianto: 6115607 Silmi Afina Aliyan: 6749474
    Uncontrolled Keywords: deformasi permukaan, DInSAR, Sentinel-1, Random Forest, Multi-Layer Perceptron, Sesar Garsela surface deformation, DInSAR, Sentinel-1, Random Forest, Multi-Layer Perceptron, Garsela Fault
    Subjects: G Geography. Anthropology. Recreation > G Geography (General)
    L Education > L Education (General)
    Q Science > Q Science (General)
    Q Science > QB Astronomy
    Divisions: Fakultas Pendidikan Ilmu Pengetahuan Sosial > Sains Informasi Geografi S1
    Depositing User: Hanipah Nurdini
    Date Deposited: 07 Nov 2025 06:49
    Last Modified: 07 Nov 2025 06:49
    URI: http://repository.upi.edu/id/eprint/144924

    Actions (login required)

    View Item View Item