.png){kind=icon}
.png){kind=icon}
Return
Spatial landslide susceptibility modelling using metaheuristic-based machine learning algorithms
.png){kind=icon}
PRE
AI
DOI：10.1007/s00366-022-01695-6.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
To prevent and mitigate landslide risks, landslide susceptibility is an essential tool to plan and manage urban development, especially in hilly regions. This study explores two metaheuristic-algorithms, i.e., artificial bee colony (ABC), and artificial fish swarm (AFS), for the optimization of artificial neural network (ANN) model. These two algorithms are integrated with ANN model to find its optimal computational parameters for landslide susceptibility mapping in the Penang Island, Malaysia. The spatial database contains twelve landslide causative factors. In this study, 382 landslide events occurred, and they are divided into two parts: two-third for train data and one-third for test data. The pre-processing technique is frequently used in machine learning approach to enhance the efficiency of a model. Therefore, the normalization and principal component analysis (PCA) are applied on the spatial database to eliminate the redundant and overlapping instances. The mean squared error (MSE), classification accuracy (Acc), and area under the receiver operating characteristic (AUROC) curve are used to evaluate the comprehensive performance of the proposed models. The obtained AUROC value of the ABC-ANN model is 96.99%, which is higher than that of AFS-ANN (96.66%) and ANN (96.43%) models. The integrated models can produce the satisfactory results for this study area. It is deduced from the obtained results that the predictive ability of ABC-ANN model is better in optimizing the computational parameters and structure of ANN model as compared to AFS-ANN. The resulting landslide susceptibility maps offer the important information for the assessment of landslide risks in this study area.
Keywords:
Landslide susceptibility
Metaheuristic-based machine learning
Artificial neural network (ANN)
Artificial bee colony (ABC)
Artificial fish swarm (AFS)
Journal
IF:
4.9
Papers:
2.6K
Citations:
9.3K