Prediction of landslide displacement based on the EEMD-GWO-LSTM
To achieve accurate prediction of landslide displacement,a fine landslide displacement time series prediction model based on the ensemble empirical modal decomposition(EEMD)data processing technique and the gray wolf optimization(GWO)algorithm long and short term memory(LSTM)network is proposed.The model uses the ensemble empirical modal decomposition method to decompose the collected cumulative landslide displacements into a number of intrinsic modal functions and one residual component,reconstructing the intrinsic modal functions into stochastic and periodic displacements,and defining the residual component as the trendline displacement.The long and short term memory network is used to predict the stochastic and periodic displacements,and the gray wolf optimization algorithm is introduced to optimize the hyper-parameters of the long and short term memory network model to establish the optimal prediction model.Then,based on the developmental characteristics of the trend displacement,the least squares method is used to realize the prediction.Finally,the accumulated displacement prediction can be obtained by summing the stochastic displacement,periodic displacement and trend line displacement.The reliability of the EEMD-GWO-LSTM model is verified by taking the Baishuihe landslide in the Three Gorges reservoir area as an example.The results show that the RSME,MAE and R2 of the model proposed in this paper are 9.82 mm,8.19 mm and 0.96,respectively,which are better than the BP neural network and the LSTM model.
Slope displacementPredictionEnsemble empirical mode decompositionGrey wolf optimizerLong and short term memory
NETL
NSTL
万方数据
