An Empirical Investigation of the Adaptive-Filtering Learning Factor
作者:
SheltonFred Ames,
期刊:
Journal of the Operational Research Society
(Taylor Available online 1987)
卷期:
Volume 38,
issue 3
页码: 269-275
ISSN:0160-5682
年代: 1987
DOI:10.1057/jors.1987.43
出版商: Taylor&Francis
关键词: analysis;computers;forecasting;methodology;microcomputers;modelling;non-linear;optimization;time series
数据来源: Taylor
摘要:
AbstractAcceptance of the adaptive-filtering parameter-estimation process has been mixed. Much of the variability between studies may result from the particular learning factor selected and may be further affected by the type of standardization used. The author presents a modification of the Makridakis and Wheelwright adaptive-filtering algorithm. This modified algorithm provides superior performance and simplifies the selection of the learning factor when used with simple autoregressive processes.
点击下载:
PDF (3309KB)
返 回