中文摘要：

介绍了小波分解法改进时间序列预报模型的原理。以IGS发布的2013年VTEC数据为例进行逐点建模,选用db4小波对5 112个格网点VTEC序列进行分解,并对各小波分量进行时间序列预报和重构,进而对VTEC进行预报。对一般方法和改进方法预报偏差的比例变化及全球范围内各格网点的预报值RMSE进行比较。结果表明,改进法预报的高精度点比例及其预报精度均优于一般方法。

英文摘要：

This paper introduces the time series prediction model of ionospheric VTEC improved by wavelet decomposition.The VTEC grid data released by IGS in 2013 are taken as experimental data to model point by point.Firstly,the VTEC series of 5 112 grid points are decomposed using db4 separately,then predicted using time series analysis,and finally reconstructed to predict the VTEC.The forecast bias and RMSE of the common prediction method and the improved method are compared.The result shows that the precision and the proportion of high accuracy using the improved method are better than the common prediction method.The improved method achieves the aim of raising accuracy.