BP神经网络算法在PM_(2.5)数据预测方面的应用

Application of BP Neural Network in PM_(2.5) Data Forecasting

针对近几年空气污染日益严重的问题,文中提出了利用BP神经网络算法预测PM_(2.5)数据的方法,选取PM_(2.5)数据样本,以及与之相关的相对湿度、空气温度、风速等级等数据样本,在MATLAB平台上构建神经网络模型,并进行训练与参数调整,建立最优的PM_(2.5)数据预测模型,实现对PM_(2.5)数据的预测。基于北京地区的实验结果表明,该方法具有良好的精度,易于工程实现,对城市空气质量预测具有实用意义。

In view of the increasingly serious problem of air pollution in recent years,in this paper,a method of predicting PM2.5data by using BP neural network algorithm is proposed.The PM2.5data samples are selected and the relative humidity,air temperature and wind speed grade are selected.The neural network model is built on MATLAB platform,training and parameter adjustment,the establishment of the optimal PM2.5 data prediction model to achieve the PM2.5data prediction.Experimental results based on the Beijing area show that the method has good accuracy and is easy to be realized in engineering.It is of practical significance for urban air quality prediction.