Displacement prediction of periodic term of Baishuihe landslide based on APSO-SVR-GRU model

Abstract

The prediction of landslide periodic term displacement is a crucial step in the study of landslide deformation in geological disasters. Since single prediction model is susceptible to accidental factors and cannot make full use of effective information, its prediction accuracy is not high and its applicability is not strong. In this paper, a combined prediction model combining adaptive particle swarm optimization (APSO), support vector machine regression (SVR) and gated neural network (GRU) algorithm is proposed. The model uses the adaptive particle swarm optimization algorithm to optimize the parameters of the support vector machine regression algorithm, determines the optimal parameter combination, and then uses the least square method to weight the APSO-SVR model and the GRU model to establish the optimal weight ratio combination model. Taking the Baishuihe landslide of the Three Gorges as the research object, selecting precipitation, reservoir water level and displacement as the influence factors of the periodic term displacement, the model is trained and verified. The results show that: in the Baishuihe landslide periodic term displacement prediction, the APSO-SVR-GRU compared with a single model has higher prediction accuracy and stability.

Keywords

• periodic term displacement
• least squares method
• svr
• gru
• apso