Echo state network (ESN) is an important method for time series prediction. However, the overfitting problem is likely to occur when the training data contain noise or outliers. To solve this problem, an ESN model based on smoothly clipped absolute deviation (SCAD) penalty function was proposed in this paper. Different from the traditional methods, such as ridge regression, L1 norm penalty, wavelet denoising and other methods added into the ESN model, the SCAD penalty function was used to select the variables of the ESN model. Specially, to meet the variable sparseness, the small coefficients are set to zero. And the large coefficients are taken as constants,which can well solve the over-fitting problem of ESN and satisfy approximate unbiased estimation. For the nonconvex optimization problem of SCAD penalty function, the local quadratic approximation (LQA) solution was presented in the paper, and the enormous computational complexity of the least angle regression (LQR) method for solving the SCAD penalty function was overcome.Then,the particle swarm optimization (PSO) is used to quickly determine the hyperparameters selection of smoothly clipped absolute deviation-echo state network (SCAD-ESN) model. The proposed method overcame the blindness of the conventional methods using the experience to select the hyperparameters, which is blind and difficult to determine the global optimum. Finally, the chaotic system simulation and network traffic simulation showed that,compared with the conventional models, the model can effectively reduce the test error and overcome overfitting problem..
A large number of landslides and collapses are often triggered by earthquake, but little attention has been paid to the risk assessment of glacier lake outburst under coaction of earthquake and ice avalanches. To investigate the characteristic of hydrodynamic pressures under coaction of earthquake and ice avalanches, experimental analysis of hydrodynamic pressures were conducted under separate earthquake action and ice avalanches action, respectively. The ice block sliding into the reservoir was added during the process of earthquake simulated by shaking table, the characteristic of combined hydrodynamic pressures caused by earthquake and ice avalanches were then observed. The results show that the seismic hydrodynamic pressures were decreased from the bottom to the top, which was in good agreement with Westergaard's briefly solutions. However the hydrodynamic pressures caused by ice avalanches were increased from the bottom to the top. Meanwhile the effect of the volume,the sliding block velocity and initial water depth were considered to study the characteristics of the hydrodynamic pressures, and an empirical equation was obtained to predict the maximum pressure loads acting on the dam. The hydrodynamic pressure distribution curves under coaction of earthquake and ice avalanches were approximately equal to the superposition of separate earthquake and ice avalanches actions,but the combined values were smaller than the superposition values, and the estimation method of maximum hydrodynamic pressure was proposed.