Prediction of Surface Displacement of Landslides Based on VMD-SegSigmoid-XGBoost-ClusterLSTM Algorithm
Predicting the surface displacement of mountain landslides can help predict new potential sliding surfaces and avoid causing more serious harm.This article conducts modeling research on the one-way surface displacement data of landslides in Niutangao Formation,Dashajie Village,Heliao Township,Zhijiang County,and proposes a time series prediction framework based on variational mode decomposition,VMD-SegSigmoid-XGBoost-ClusterLSTM,which can accurately predict the surface displacement of landslides.The model performs well in the data set,the root-mean-square deviation and average absolute percentage error of trend subsequence and periodic subsequence are both less than 0.1,except for the residual subsequence that is difficult to fit,the root-mean-square deviation of XGBoost periodic prediction module is as low as 0.006.
万方数据
维普
