Research Article
Random Forest Method Utilization For Landslide Hazard Zonation In Lima Puluh Kota Regency
@INPROCEEDINGS{10.4108/eai.19-11-2022.2332288, author={Ahyuni Ahyuni and Rizki Atthoriq Hidayat and Endah Purwaningsih and Yurni Suasti}, title={Random Forest Method Utilization For Landslide Hazard Zonation In Lima Puluh Kota Regency}, proceedings={Proceedings of the 2nd International Conference of Geography, ICGEO 2022, 19-20 November 2022, Padang, West Sumatera, Indonesia}, publisher={EAI}, proceedings_a={ICGEO}, year={2023}, month={6}, keywords={landslide; random forest; hazard zonation}, doi={10.4108/eai.19-11-2022.2332288} }- Ahyuni Ahyuni
Rizki Atthoriq Hidayat
Endah Purwaningsih
Yurni Suasti
Year: 2023
Random Forest Method Utilization For Landslide Hazard Zonation In Lima Puluh Kota Regency
ICGEO
EAI
DOI: 10.4108/eai.19-11-2022.2332288
Abstract
There are various statistical methods that are applicable for landslide hazard zonation. Random Forest classification method is deployed in determining landslide hazard in Lima Puluh Kota Regency. Landslide records are obtained from field measurement for 145 points, 80% of total points used as train data and 20% for test data. Predictor data obtained from related agencies, namely rainfall, geology, LULC, soil type, aspect, elevation, slope, curvature, road distance, and river distance. Random Forest analysis applied 1000 trees to get probability values. ROC was obtained for 0.8137 which represents the model is qualified to predict landslide hazard zones at very high (6.1% ), high (11.106%), moderate (19.036%), and low (63.748%).