BP神经网络在北京市API预报中的应用

对大气污染进行预测具有十分重要的意义。以北京市为例,综合考虑BP网络的逼近能力和泛化能力,将时间序列作为BP神经网络的输入,对空气污染指数的预测做了建模研究。实验结果表明:BP神经网络的输出值与实际值之间的误差在可以接受的范围,特别是对骤升骤降趋势也能得到准确度较高的预报结果。因此在实际应用中,可以将BP网络方法作为一种考虑采用的方法。

APLICATION OF BP NEURAL NETWORK TO FORECASTING API IN BEIJING

Forecasting the level of air pollution has a great significance.Taking Beijing as an example,the optimal network structure is determined by making an integrated survey of the approximation capability and the generalization of the network.The time series is taken as the input of BP neural network,and a lot of research has been done on forecasting the level of air pollution.The experimental result indicates that the error between the output of the neural network and the actual numerical values is in the acceptable range.It can be gained a highly precise forecast result in the trend of increasing dramatically and declining sharply.As a result,the network can be considered as an adoptable method in practice.