A novel hybrid integration model using support vector machines and random subspace for weather-triggered landslide susceptibility assessment in the Wuning area (China)

Abstract

This study proposed a hybrid modelling approach using two methods: support vector machines and random subspace to create a novel model namely Random SubSpace based Support Vector Machines (RSSVM) for landslide susceptibility assessment. The newly developed model was then tested at Wuning area, China to produce landslide susceptibility map. To achieve the objective of the study, first a spatial dataset was constructed that includes a landslide inventory map consisting of 445 landslide locations; and various landslide conditioning factors, such as slope, aspect, altitude, topographic wetness index (TWI), stream power index (SPI), sediment transport index (STI), soil, lithology, normalized difference vegetation index (NDVI), land use, rainfall, distance to roads, distance to rivers, distance to faults. Next, the result of RSSVM model was validated using statistical index-based evaluations and receiver operating characteristic curve method. Next, to assess the performance of the proposed RSSVM model, a comparison analysis was performed with other existing methods such as ANN (Artificial Neural Network), NB (Naïve Bayes) and SVM (Support Vector Machine). In general, the performance of the RSSVM model showed higher than the rest of the models for spatial prediction of landslides. The AUC results of the applied models are: RSSVM (AUC=0.857), followed by MLP (AUC=0.823), SVM model (AUC=0.814) and NB model (AUC=0.783). The present study indicates that RSSVM can be used for spatial prediction of landslides and the result is very useful for the lcoal government and people lived in the Wuning area.

Sinkhole susceptibility mapping: A comparison between Bayes‐based machine learning algorithms

“…The area under the ROC curve (AUROC) is a measure of the capability of the model to predict the spatial distribution of events (sinkhole; Hong et al, ). It ranges between 0.5 (null prediction capability) and 1 (perfect model; Shirzadi, Bui, et al, ).…”

“…In conditional independence, it is assumed that all attributes of examples are independent for maximizing the posterior probability with given output class for classification issue (Soni, Ansari, Sharma, & Soni, 2011). The main aim of the NB classifier is to compute the prior probabilities of each class using a discriminant function (Hong et al, 2017). NB has been applied in many scientific fields because it is very robust to noise and irrelevant attributes and also does not need a big training dataset for modeling (Tien Bui, Pradhan, Lofman, & Revhaug, 2012).…”

“…Additionally, the MLAs can deal with spatial patterns of data at various scales (Kanevski et al, 2004). The application of these new machine learning predictive models has been explored in different geoscience fields including landsides (Chen, Niu, Du, & Wang, 2015;Chen, Niu, & Jia, 2016;Chen, Trinder, & Niu, 2017;Hong et al, 2017), groundwater qanat potential (Naghibi, Pourghasemi, & Abbaspour, 2017), or land subsidence (Pradhan, Abokharima, Jebur, & Tehrany, 2014). However, their application to sinkhole susceptibility modeling is still very limited.…”

Landslide Susceptibility Evaluation Using Hybrid Integration of Evidential Belief Function and Machine Learning Techniques

“…In addition, some studies have shown that RS is an excellent ensemble algorithm. Hong et al indicate that the integrated model RSSVM has the optimal performance [55]. In this study, training (goodness of fit) and validation (prediction accuracy) data sets were used to compare the proposed integration model with the benchmark model.…”

“…The mixed model RSSVM model also achieved optimal performance (AUC = 0.857) [55]. Therefore, the hybrid models are often considered to be more precise to single models [55].The Loess Plateau is one of the most vulnerable regions in China. The special geological environment of the area has led to the development of environmental geological problems.…”