EMD-SVM在南京市月平均气温预测中的应用

南京市月平均气温具有非平稳性、噪声大、序列宽频等特征.为了提高温预测精度,本文提出一种经验模态分解(EMD)和支持向量机(SVM)回归相组合的预测模型(EMD-SVM).首先应用EMD分解算法把南京市月平均气温分解成不同尺度的基本模态分量(IMF),再运用支持向量机回归模型对每个IMF预测,最后将预测结果重构得到南京市月平均气温预测值.结果表明:EMD-SVM模型预测与单一支持向量机回归模型预测相比,平均预测精度提高0.59度,是一种有效的预测气温的模型.

Nanjing Monthly Average Temperature Prediction Base on Empirical Mode Decomposition and Support Vector Machine

The monthly average temperature of Nanjing has some characteristics,such as non-stationary,great noise,sequence-wideband and so on.In order to improve the prediction accuracy of the Temperature in Nanjing,based on empirical mode decomposition(EMD),a temperature prediction model(EMD-SVM) is put forward by support vector machine regression(SVM).Firstly,the time serial of the monthly average temperature in Nanjing is decomposed into many intrinsic model functions(IMF) by using EMD.Secondly,we utilize SVM to predict every IMF.Finally,we can get the predicted value of the monthly average temperature in Nanjing by reconstructing the forecasting results.The result has shown that the EMDSVM is a higher and more efficient temperature prediction model,owing to comparing with the SVM,the precision of EMD-SVM model is increased by 0.59 degrees.