A stepwise procedure for the selection of nonlinear regression models
作者:
H. Bunke,
B. Droge,
期刊:
Statistics
(Taylor Available online 1985)
卷期:
Volume 16,
issue 1
页码: 35-45
ISSN:0233-1888
年代: 1985
DOI:10.1080/02331888508801822
出版商: Akademie-Verlag
关键词: Model selection;stepwise procedure;nonlinear regressin;computer programme
数据来源: Taylor
摘要:
A stepwise procedure is proposed for the selection of nonlinear regression models which, in some sense, compromises between the sum of squared residuals and the parameter dimension, trying to keep this dimension as low as possible. The procedure is based on asymptotic results. For the case of univariate regressin, a computer programme has been implemented. The problem of choosing the class of admitted models is also discussed
点击下载:
PDF (3142KB)
返 回