nlout: Nonlinear outlier detection.

In nlr: Nonlinear Regression Modelling using Robust Methods

Description

Detecting outlier for nonlinear regression, is based on mixing robust estimates and statsitics measures.

Usage

nlout(nlfited)

Arguments

nlfited

Object of type nl.fitt or nl.fitt.gn for classic estimators, nl.fitt.rob or nl.fitt.rgn for robust estimators.

Details

The outlier detection measutred used in this function are studentized residuals and Cook Distance. They are mixture of estimators and Jacobians. They are successful for detecting outlier only if combine with robust fits, eventhough the function can work with classic fits but it is not recomended.

Value

Result is list of nl.robmeas objects for each statistics.

"vmat"	variance covariance matrix of parameters σ^2 (\nabla f(θ)'\nabla f(θ))^{-1})
"d.yhat"	predicted values after rremoving a point \hat y_{(-i)}
"studres"	nl.robmeas object Studentized residuals.
"cook"	nl.robmeas object od Elliptic Norm (Cook Dist)
"mahd.v"	nl.robmeas object of Regression Mahalanobis Distance.
"mahd.dt"	nl.robmeas object of Mahalanobis MVE, data.
"mahd.xs"	nl.robmeas object of Mahalanobis MVE, xs.
"hadi"	nl.robmeas object of Hadi potential.
"potmah"	nl.robmeas object of Potential mahalanobis.
"delstud"	nl.robmeas object of Deletion Studentized.
"dffits"	nl.robmeas object of DFFITS.
"atk"	nl.robmeas object of Atkinson Distance.
"mvedta"	nl.robmeas object of MVE data.
"mvex"	nl.robmeas object of MVE x.
"dfbetas"	nl.robmeas object of DFBETAS.

Note

This function return back all resutls and statistics but, Riazoshams (2009) showed studentized residuals and Cook distance when combine with robust estimators can detect outliers correctly. Thus to identify outlier correctly first estimate the parameters bu robust options of nlr function then call nlout, finally look at the list values "studres" and "cook" from the result list. The plot and other methods of nl.robmeas display the results visually.

Author(s)

Hossein Riazoshams, Dec 2008 Email: [e-mail] URL http://www.riazoshams.com/nlr/

References

Riazoshams H, Habshah M and Adam MB 2009 On the outlier detection in nonlinear regression. 3(12), 243-250.

See Also

nl.fitt, nl.fitt.gn, nl.fitt, nl.fitt.gn, nl.fitt.rob, nl.fitt.rgn, nl.robmeas, nlr, nlout.JL

Examples

 d<-list(xr=Weights$Date, yr=Weights$Weight)
 wmodel <- nlr(nlrobj1[[2]],data=d,control=nlr.control(method = "OLS",trace=TRUE))
 a=nlout(wmodel)
 ## Run the command as bellow
 ## nlout(wmodel)

nlr documentation built on July 31, 2019, 5:09 p.m.