水资源与水工程学报

郭彦，林秀芝，侯素珍，蒋尚明，景林艳.基于EMD和BP算法的降水数据插补水资源与水工程学报[J].,2015,26(2):16-21

基于EMD和BP算法的降水数据插补

Interpolation of precipitation data based on empirical mode decomposition and back propagation algorithm

DOI：10.11705/j.issn.1672-643X.2015.02.004

中文关键词: 降水资料插补延长经验模态分解 BP算法逐步回归分析频率分析相关分析

英文关键词:precipitation data interpolation and extrapolation empirical mode decomposition back propagation algorithm stepwise regression analysis frequency analysis correlation analysis

基金项目:“十二五”国家科技支撑计划项目（2012BAB02B03）；中央级公益性科研院所基本科研业务费专项（HKY-JBYW-2014-05）；国家重点基础研究发展计划（973计划）

作者	单位
郭彦^1,2，林秀芝^1,2，侯素珍^1,2，蒋尚明³，景林艳⁴	(1.黄河水利委员会黄河水利科学研究院，河南郑州 450003； 2.水利部黄河泥沙重点实验室，河南郑州 450003；3.安徽省·水利部淮河水利委员会水利科学研究院安徽省水利水资源重点实验室，安徽蚌埠 233000；4.河南省林州市水务局，河南林州 456550)

摘要点击次数: 1926

全文下载次数: 1037

中文摘要:

降水数据的插补可有效地改善数据系列的完整性。以西柳沟流域作为毛不拉孔兑的参证流域，采用EDM和BP算法对毛不拉孔兑缺失的降水量进行插补，并与已有的相关分析法、频率分析法及逐步回归分析法进行了比较。研究结果表明：EDM-BP模型综合利用了西柳沟降水序列EMD分量的有效信息，因而能较好地反映降水量的变化趋势，在数据拟合和预测方面达到较高的精度，适合毛不拉孔兑降水数据的插补，在其它水文数据插补延长中具有一定的推广应用价值。

英文摘要:

The interpolation of precipitation data can improve the integrity of data series effectively. Taking Xiliugou River basin as reference, the paper interpolated the missing precipitation data in Maobulakongtuo interpolated with Empirical Mode Decomposition (EMD) and Back Propagation algorithm (BP) methods.The results were compared with that of correlation analysis, frequency analysis and stepwise regression analysis. The results show that EDM-BP model takes full advantage of effective information of Xiliugou the interpolation of precipitation data in EDM decomposition series,therefore which can better reflect the trend of precipitation series. Its precise of data fit and forecast is higher and is suit to the interpolation of precipitation data at Maobulakongtuo.The result has certain application value to the interpolation and extrapolation of other hydrological data.

查看全文查看/发表评论下载PDF阅读器

关闭