基于小波与径向基函数的大气污染预测研究

提出了一种实用于大气污染物的小波分析和RBF(RadialBasisFunction)神经网络相结合的预测方法,对污染物TSP(TotalSuspendedParticles)进行预测,得到与观测值相符的结果。应用小波分析进行检测,消除了预测数据中奇异点所包含的奇异信号。在此基础上应用RBF神经网络进行预测,研究结果表明,小波分析的利用提高了预测精度,验证了该方法简便、快捷,具有较强的实用性,为环境管理部门宏观调控提供了理论基础。

The Research and Prediction of Atmospheric Pollution Based on Wavelet and RBF Neural Network

Based on the wavelet analysis and the RBF(Radial Basis Function)neural network,a predicting method is proposed,which is applied to protect the air contamination TSP(Total Suspended Particles).Firstly,the wavelet analysis is utilized to check and remove the strange signal in the strange point of predicted data,which may increase the predictive precision.Secondly,the RBF neural network is applied to predict the air contamination TSP.Finally,an example is included to verify the precision and effectiveness of the proposal method,and the result indicates that the predicted values agree with the observation data.Meanwhile,it is shown that this method is simple and fast in various cases successfully.Above all,this method has the stronger practicability and provides the theory foundation for macroscopical adjusting of the environment management.