Hakim, W. L., Rezaie, F., Nur, A. S., Panahi, M., Khosravi, K., Lee, C. W., & Lee, S. (2022). Convolutional neural network (CNN) with metaheuristic optimization algorithms for landslide susceptibility mapping in Icheon, South Korea. Journal of environmental management, 305, 114367. Click (a) CNN-GWO based landslide susceptibility map: Spatial distribution of landslide risk across Icheon City, derived through the optimized deep learning model (CNN-GWO). (b) Assessment of variable importance (IGR): Results of the Information Gain Ratio (IGR) analysis, identifying forest diameter, forest age, and forest type as the most critical geo-environmental factors for landslide occurrence. (c) ROC-AUC performance metrics: Validation graph demonstrating that the proposed CNN-GWO model achieved an AUC of 0.876, securing superior predictive performance over the standalone CNN and CNN-ICA models. This study investigates the application of deep learning algorithms based on Convolutional Neural Networks (CNNs) coupled with metaheuristic optimization algorithms to generate highly accurate landslide susceptibility maps for Icheon City, South Korea. A total of 18 geo-environmental and topo-hydrological factors were utilized as predictive variables, and the existing landslide inventory dataset was randomly divided into 70% for training and 30% for validation.The results indicated that integrating the Grey Wolf Optimizer (GWO) and the Imperialist Competitive Algorithm (ICA) successfully improved the spatial prediction accuracy compared to the standalone CNN model (AUC = 0.847, RMSE = 0.12). The CNN-GWO model exhibited the highest predictive capability, achieving an AUC of 0.876 and an RMSE of 0.08. Information Gain Ratio assessment revealed that forest diameter, forest age, and forest type were the most crucial contributing factors to landslide occurrence in the study area. The proposed optimized deep learning methodology proves to be a robust and reliable tool for predicting landslide-prone areas, assisting local governments in disaster mitigation and sustainable development planning. Continue reading...
