Abstract：The neural network method is a powerful tool to solve the problem of nonlinearity，but when the input variables are very large,the multicollinearity among variables will lead low modeling efficiency.Partial least-squares regression(PLSR) method can extract compositions with better interpretation to the dependent variables,and it can preferably overcome the multicollinearity among variables. By combining the two methods, a BP neural network prediction model for peak velocity of blasting vibration based on PLSR was established. The affecting factors for blasting vibration were analyzed by means of PLSR,and three new synthesis variables were extracted. Since input layers of BP neural network were decreased from nine to three,the network structure is simpler,more efficient and more stable.The results show that the average prediction errors of coupled model is 7.62%,which is more accurate than Sadaovsk formula and normal BP neural network model.
史秀志;武永猛;唐礼忠;黄宣东. 偏最小二乘回归神经网络模型在爆破振动峰值速度预测中的应用[J]. , 2013, 32(12): 45-49.
SHI Xiu-zhi;WU Yong-meng;TANG Li-zhong;HUANG Xuan-dong. Application of neural network model with partial least-squares regression in prediction of peak velocity of blasting vibration. , 2013, 32(12): 45-49.
2013, Vol. 32 
