Nonparametric Regression with Correlated Errors

Simple item page
dc.contributor.authorJose, C.T.
dc.contributor.authorIsmail, B.
dc.date.accessioned2014-07-08T04:43:50Z
dc.date.available2014-07-08T04:43:50Z
dc.date.issued2007-02
dc.description.abstractLinear smoothing is a popular technique in estimating the mean function in a nonparametric regression model y=m(x)+s, where m(x) is a smooth function and e is an iid error with mean zero. The linear smoothing technique is extended to accommodate a correlated error process. The cross-validation criterion for choosing the optimum bandwidth performs very badly when the errors are correlated. A method is proposed to estimate the error covariance function based on the residuals from a linear regression smoother. Using the estimated covariance function, the regression model is transformed to produce uncorrelated transformed errors. The nonparametric regression function estimate is obtained by using the linear smoothing technique on the transformed model. The method is illustrated through simulation studies.en_US
dc.identifier.citation53rd Annual Conference of Indian Society of Agricultural Statistics, Tiruchirappalli (2 - 4 Dec, 1999)en_US
dc.identifier.urihttp://hdl.handle.net/123456789/3818
dc.language.isoenen_US
dc.subjectBandwidth selectionen_US
dc.subjectcorrelated errorsen_US
dc.subjectcovariance functionen_US
dc.subjectcross-validationen_US
dc.subjectnonparametric regressionen_US
dc.titleNonparametric Regression with Correlated Errorsen_US
dc.typeArticleen_US

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
nonparametric regression with correlated errors.pdf
Size:
443.84 KB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Publications