Utilization of Satellite Imagery and Machine Learning for Mapping Landslide-Prone Areas in Banjarnegara Regency, Central Java Province

The high risk of landslides in Banjarnegara Regency requires preparedness for the potential for landslides. One way that can be utilized by the government and the community is to use a mitigation map that illustrates the vulnerability of landslides in Banjarnegara Regency. This study aims to produce a landslide sus-ceptibility map with satellite imagery and the best machine learning model. The variables used in this study include rainfall, soil type, geology, aspect, curvature, slope gradient, elevation, TWI, NDWI, NDVI, NDBI, and land cover. Based on the results obtained using the Random Forest, Support Vector Machine, Logistic Regression, and Artificial Neural Network models, in 2021 and 2022, the Random Forest model had the best performance with an accuracy of 82.50% and 88.46%, respectively, while in 2022 the model with the best performance was the Support Vector Machine with an accuracy of 86.84%.