Prediction of the landslide susceptibility: Which algorithm, which precision?

• Published in: Catena (Giessen) Vol. 162; pp. 177 - 192
• Main Authors: Pourghasemi, Hamid Reza, Rahmati, Omid
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2018
• Subjects: analytical methods; computer software; discriminant analysis; geographic information systems; GIS; Iran; Landslide spatial modeling; landslides; linear models; Machine learning; neural networks; prediction; regression analysis; support vector machines; Variables importance; Iran; GIS; Variables importance; Landslide spatial modeling; Iran; Machine learning
• ISSN: 0341-8162; 1872-6887
• DOI: 10.1016/j.catena.2017.11.022

Coupling machine learning algorithms with spatial analytical techniques for landslide susceptibility modeling is a worth considering issue. So, the current research intend to present the first comprehensive comparison among the performances of ten advanced machine learning techniques (MLTs) including artificial neural networks (ANNs), boosted regression tree (BRT), classification and regression trees (CART), generalized linear model (GLM), generalized additive model (GAM), multivariate adaptive regression splines (MARS), naïve Bayes (NB), quadratic discriminant analysis (QDA), random forest (RF), and support vector machines (SVM) for modeling landslide susceptibility and evaluating the importance of variables in GIS and R open source software. This study was carried out in the Ghaemshahr Region, Iran. The performance of MLTs has been evaluated using the area under ROC curve (AUC-ROC) approach. The results showed that AUC values for ten MLTs vary from 62.4 to 83.7%. It has been found that the RF (AUC=83.7%) and BRT (AUC=80.7%) have the best performances comparison to other MLTs. [Display omitted] •Landslide susceptibility was assessed using different geo-environmental factors.•The study compared the accuracy of 10 advanced machine learning techniques.•Learning vector quantization algorithm determined the variables' importance.•Partial response curves were used to understand the role of conditioning factors.•The models' performances were assessed using the AUC-ROC method.