关键词: B-P神经网络, 空气质量, 预测模型, 隐层

Abstract: B-P neural network is a powerful tool that describes nonlinear phenomenon.We can apply it into the forecasting work of the environment air quality.For the different monitoring item it will make several groups of training data to train the B-P neural network and make it learn in light of the different weather character on the basis of taking the pollute source exhausting data as the input data and taking the monitoring data in monitor position as the output data.And it will establish different prediction network.And then we can output the corresponding item’s monitoring data in the monitor position by inputting the emitting monitor data of the air pollution resource into the trained B-P neural network with the adjusted height in the same weather condition.The experiment proves that B-P neural network prediction model has acquired much better result and it has much more superiority than the current prediction model.

Key words: B-P neural network, air quality, forecasting model, hidden layer