Research On The Principal Component Analysis-BP Neural Network In The Prediction Of Coal Mine Cas Emission Amount

China is the largest coal manufacturing and consuming country in the world. However, the coal mining industry is of high risk that gas emission can bring about suffocation, even explosion in the process of coal mining, thus pose a threat to coal mine safety in production. Therefore, it is an issue of great importance to accurate the prediction of the amount of gas emission to ensure the safety in production. Many scholars have done a lot work on this. The prediction of mine-gas emission based on neural network is widely used due to its characters such as self-organizing, self-adaptive and parallel-processing. However, as the new technology and new equipment join in the coal mine production system, there are more and more gas emission prediction factors we should consider. But the too big dimensions of the input data can lead to a surge in the network scale that can reduce the generalization ability and convergence of the network. Most of existing studies only consider part of the factors and often ignore the correlation between influence factors, so their prediction precision is difficult to guarantee.This thesis brings forward the BP neural network based on principal component analysis, which aimed at the intricate issues of mine-gas emission prediction, using principal component analysis to deal with multivariable parameter matrix, and we could get less principal components reflecting the original variables, thus simplifies the original multidimensional problems. The principal components obtained can be the input for the prediction of mine-gas emission based on BP neural network. In addition, we conducted the experiment to compare the principal component analysis based on BP neural network with the traditional BP neural network and RBF neural network. As a result, the experiment proved that the principal component analysis based on BP neural network did well.