冲击地压的发生会对矿山造成巨大的经济损失,预测预报是防治冲击地压的重要组成部分。文章将主成分分析方法与反向传播人工神经网络(即BP神经网络)方法相结合,用于对冲击地压的预测。首先,利用主成分分析法对冲击地压的影响因素进行分析,降低数据的维数。然后,将所得的数据作为BP神经网络的输人数据进行训练,用训练好的神经网络对冲击地压进行预测。预测结果与实际结果的对比表明,利用主成分分析一BP神经网络方法对冲击地压预测是可行的。
The occurrence of rock burst could cause enormous economic loss to the mine enterpri ses. Forecasting plays an important role in preventing rock burst. This paper uses principle compo nent analysis combined with backpropagation artificial neural network (BP neural network) for the prediction of rock burst. Firstly, principle component analysis is used to analyse the influential fac tors and reduce the dimensions of data samples. Then, the calculated results are used as input data for training the neural network, and the trained network is used to predict rock burst. Comparison of forecast and actual results suggests that it is feasible to forecast rock burst by using principal compo nent analysis and BP neural network method.