A Novel Decomposition-Ensemble Learning Model Based on Ensemble Empirical Mode Decomposition and Recurrent Neural Network for Landslide Displacement Prediction

As vital comments on landslide early warning systems, accurate and reliable displacement prediction is essential of significant importance for mitigation. However, obtaining the desired accuracy remains highly difficult challenging due to complex nonlinear characteristics monitoring data. Based principle “decomposition ensemble”, a three-step decomposition-ensemble learning model integrating ensemble empirical mode decomposition (EEMD) recurrent neural network (RNN) was proposed prediction. EEMD kurtosis criteria were first applied data construction trend periodic components. Second, polynomial regression RNN with maximal information coefficient (MIC)-based input variable selection implemented individual components independently. Finally, predictions aggregated into final The experimental results from Muyubao demonstrate that EEMD-RNN capable increasing outperforms traditional models (including EEMD-support vector machine, EEMD-extreme machine). Moreover, compared standard RNN, gated unit (GRU)-and long short-term memory (LSTM)-based perform better in predicting accuracy. promising