Abstract
Abstract. Java Island is also an island with a high frequency of landslide natural disaster. Various efforts have been made to minimize the disaster risk, including the model compilation of early disaster detection at the areas with landslide potential. This research aims to develop an early warning model for landslide potential areas using Spatial Autocorrelation combined with Machine Learning Algorithm based on attributes of landslide causes. The first step is to classify areas in Java Island into a landslide hotspot and landslide coldspot using spatial autocorrelation G * algorithm. This algorithm revealed 124 polygons of sub-district in Java as landslide hotspots. The next step is building machine learning model using C5.0 method for hotspot and coldspot areas. In this research, we utilize landslide-causing attributes i.e. rainfall, land cover, area slope, soil type, and land movement. The hotspot model showed that landslide distribution focuses on land cover attributes. Meanwhile, for the coldspot area model, there is no focus of landslide distribution on one attribute. Furthermore, the accuracy level of hotspot model are 84.61% and 71.66% for coldspot model. Keywords: Landslide, Machine Learning, Spatial Autocorrelation, C5.0, G*.Abstrak. Pulau Jawa merupakan pulau dengan frekuensi bencana alam tanah longsor yang tinggi. Untuk meminimalkan resiko bencana, dilakukan penyusunan model komputasi deteksi dini daerah potensi longsor. Tujuan dari penelitian ini adalah membangun model deteksi daerah potensi longsor menggunakan Spatial Autocorrelation yang dikombinasikan dengan algoritma Machine Learning berbasis data variabel-variabel pemicu longsor. Tahap pertama yang dilakukan adalah mengklasifikasikan daerah-daerah di Pulau Jawa sebagai daerah hotspot longsor dan coldspot longsor menggunakan algoritma spatial autocorrelation G*. Dihasilkan 124 poligon kecamatan di pulau Jawa sebagai daerah hotspot longsor. Setelah diklasifikasikan, dibangun model machine learning menggunakan metode C5.0. Atribut-atribut pemicu longsor yang digunakan untuk membangun machine learning adalah curah hujan, tutupan lahan, kelerengan, jenis tanah, dan gerakan tanah. Hasil yang diperoleh dari model hotspot terlihat distribusi longsor memusat pada atribut tutupan lahan dan menghasilkan akurasi model 84.61%. Untuk model coldspot area tidak terlihat adanya pemusatan pada satu atribut pemicu longsor, akurasi untuk model ini adalah 71.66%. Kata Kunci: Longsor, Machine Learning, Spatial Autocorrelation, C5.0, G*.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.