摘要
利用MATLAB工具箱,采用主成分分析(PCA)与径向基函数(RBF)神经网络结合,构建广西一区(北部)五月的逐日降水预报模型。在2008年5月31天独立样本的预测检验中,预测的均方误差、最大绝对误差及平均绝对误差均低于同期的T213(中国气象局的全球中期天气数值预报产品)模式,预报准确度更高。
Uses MATLAB toolbox, establishes a daily rainfall forecast model of northern Guangxi in May by using principal component analysis and radial basis function neural network. In 31 days independent sample test of the prediction in May 2008, the maximum absolute error,squared error and mean absolute error of predictionare are all lower than the same period of T213（China Meteorological Administration,the global medium range weather numerical forecast products）. The model has higher forecasting accuracy.
出处
《现代计算机》
2015年第12期7-10,共4页
Modern Computer
基金
广西高校科学技术研究项目(No.LX2014495)
关键词
主成分分析
RBF神经网络
逐日降水预测
Principal Component Analysis
RBF Neural Network
Daily Rainfall Forecast
