Using a kernel extreme learning machine with grey wolf optimization to predict the displacement of step-like landslide

Abstract

Landslide displacement prediction is an important aspect of landslide hazard research. In this paper we assess the characteristics of landslide deformation in the Three Gorges Reservoir Area of China and propose and apply a step-like displacement prediction model based on a kernel extreme learning machine with grey wolf optimization (GWO-KELM) to the Baishuihe landslide. In this model, the cumulative displacement is first decomposed into trend displacement and periodic displacement by time series. The trend displacement is then predicted by a cubic polynomial model, and the periodic displacement is predicted by the proposed model after the displacement data have been statistically analyzed. A hybrid model is then established for the prediction of landslide displacement. We then compare the performance of this hybrid model with that of the extreme learning machine with GWO (GWO-ELM), support vector machine with GWO (GWO-SVM) and extreme learning machine (ELM) models. The results show that the proposed hybrid model outperforms the other models and that the GWO-KELM model achieves excellent performance in predicting landslide displacement with a step-like behavior.