Coupling Progressive Deep Learning with the AdaBoost Framework for Landslide Displacement Rate Prediction in the Baihetan Dam Reservoir, China

• DOI: 10.3390/rs15092296
• Publication Date: 2023-04-27T08:30:47Z
• Authors: Weida Ni; Liuyuan Zhao; Lele Zhang; Ke Xing; Jie Dou
• Keywords: AdaBoost; Ensemble Learning

Disasters caused by landslides pose a considerable threat to people’s lives and property, resulting in substantial losses each year. Landslide displacement rate prediction (LDRP) provides useful fundamental tool for mitigating landslide disasters. However, more accurately predicting LDRP remains challenge the study of landslides. Lately, ensemble deep learning algorithms have shown promise delivering precise effective spatial modeling solution. The core aims this research are explore evaluate capability three progressive evolutionary (DL) techniques, i.e., recurrent neural network (RNN), long short-term memory (LSTM), gated unit (GRU) AdaBoost algorithm rainfall-induced reservoir-induced Baihetan reservoir area China. outcomes show that DL model could predict Wangjiashan with improved accuracy. highest accuracy was achieved testing set when window length equaled 30. assembling two predictors outperformed predictors, mean absolute error root square reaching 1.019 1.300, respectively. These findings suggest combination strong learners can yield satisfactory results.