Comparison of the frequency ratio, index of entropy, and artificial neural networks methods for landslide susceptibility mapping: A case study in Pınarbaşı/Kastamonu (North of Turkey)
Küçük Resim Yok
Tarih
2021
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Elsevier
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
The selection of a suitable method is crucial for landslide susceptibility mapping (LSM). The main objective of this article was to compare the index of entropy (IoE), frequency ratio (FR), and artificial neural network (ANN) methods utilized in LSM. Landslide conditioning factors such as slope, distance to roads, aspect, curvature, plan curvature, elevation, profile curvature, distance to streams, soil types, topographic wetness index (TWI), and lithology have been used to carry out the LSM of the Pınarbaşı district (Kastamonu-Turkey). All models were compared considering their prediction rates obtained using the Area Under Curve (AUC) method. The models were evaluated with a total of 1000 points, including landslide and non-landslide areas. The findings of this study show that the AUC accuracy of the FR, IoE and ANN models were 0.873, 0.869, and 0.962, correspondingly. The ANN model achieved the highest accuracy. The AUC of both the FR and IoE models showed reasonably good accuracy for producing a landslide susceptibility map. The FR and IoE methods are straightforward and easy to implement compared to ANNs. Therefore, both can be efficiently used for the LSM. © 2022 Elsevier Inc. All rights reserved.
Açıklama
Anahtar Kelimeler
Artificial neural networks, Comparison, Frequency ratio, Index of entropy, Landslide susceptibility mapping, Turkey
Kaynak
Computers in Earth and Environmental Sciences: Artificial Intelligence and Advanced Technologies in Hazards and Risk Management
WoS Q Değeri
Scopus Q Değeri
N/A
