小波时间序列在电离层TEC预报中的应用

摘要

将小波时间序列应用到电离层TEC的预测中,通过小波分解将逐天的TEC数据分解到小波频率通道上,得到对应于小波频率的小波系数列.对系数列作降噪处理并进行数学变换最终使数据趋于平稳,并进行时间序列模型预测,还原成系数列的预测值,最后还原成实际预测值.采用小波时间序列预测后5天480个样本点,准确率在80％以上占87％,90％以上占57％,相较于传统预测模型有较明显的提高.%The wavelet analysis and times series were applied for ionosphere TEC forecasting,TEC data was decomposed on wavelet frequency channel by wavelet decomposition,the discrete wavelet coefficients correspond to wavelet frequency were obtained. De-noise processing and mathematical transformation was used to get stable data. By time series prediction,predictive value was got. Finally,practical predicting results were obtained. Five days with 480 sample points were predicted in wavelet analysis and times series,the accuracies with 80% and 90% were accounted for 87% and 57% .respectively,compared to traditional prediction model,the accuracy was greatly improved.