摘要
本文介绍了均值生成函数这一时间序列分析中的新概念,并经过主成分分析建立了时间序列模拟与预测的数学模型(记作MMFA)。该模型应用于黄河下游利津河段封河日期序列的模拟和预报。结果表明,1950-1981年度序列拟合的平均相对误差为9.5%,1981-1991年度序列预报的精度达到85%以上。总之,MMFA模型应用于黄河下游封冻日期长期预报是完全可行的。同时指出了MMFA模型的实用性和存在的不足之处。
Mean Generation Function───A new concept in time series analysis is employed in thispaper, and an MMFA model about main factor analysis is presented too. It is showed thatMMFA model can simulate freeze-up dates in time series on Lijin reach of the Yellow River. and the simulated error and predicted precision are 9.5% and 85% respectively. The modelperformance and its wide applicability are presented at the end of this paper.
出处
《冰川冻土》
CSCD
北大核心
1994年第1期53-59,共7页
Journal of Glaciology and Geocryology
关键词
河流
封冻日期
长期预报
数字模型
: mean generation function, main factor analysis and modelling in time series,date prediction of river freeze-up