A Hybrid Machine Learning Model Coupling Double Exponential Smoothing and ELM to Predict Multi-Factor Landslide Displacement

The deformation of landslides is a non-linear dynamic and complex process due to the impacts both inherent external factors. Understanding basis landslide essential prevent damage properties losses life. To forecast displacement, hybrid machine learning model proposed, in which Variational Modal Decomposition (VMD) implemented decompose measured total surface displacement into trend periodic components. Double Exponential Smoothing algorithm (DES) Extreme Learning Machine (ELM) were adopted predict respectively. Particle Swarm Optimization (PSO) was selected obtain optimal ELM model. proposed method implementation procedures illustrated by step-like Three Gorges Reservoir area. For comparison, Least Square Support Vector (LSSVM) Convolutional Neutral Network–Gated Recurrent Unit (CNN–GRU) also conducted with same dataset component. application results show that DES-PSO-ELM outperformed other two methods prediction, RMSE, MAE, MAPE, R2 values 1.295mm, 0.998 mm, 0.008%, 0.999,