nl.fitt.rgn: Class '"nl.fitt.rgn"'

In nlr: Nonlinear Regression Modelling using Robust Methods

Description

Object of generalized robust estimates of nonlinear regression model.

Objects from the Class

Objects can be created by calls of the form new("nl.fitt.rgn", ...).

Slots

Robust generalized slots:

vm:

Object of class "matrix" of variance covariance matrix of error.

rm:

Object of class "matrix" of correlated error.

hetro:

Object of class "nl.fittorNULL", include object nl.fitt of heteroscedastic error fit, or NULL for non hetroscedastic. It include parameter estimates of hetroscedastic variance τ and all other slots of nl.fitt object which represent the variance function fitt information.

autcorr:

Object of class "listorNULL" of autocorrelated error.

autpar:

Object of class "listorNULL" of aprameters for autocorrelated error.

gresponse:

Object of class "vectororMatrix" generalized response, transformed response equal R \times y, for cholesky decomposition R of covariance matrix of error.

gpredictor:

Object of class "vectororMatrix" generalized predictor, transformed of predictor equal R \times η(θ), for cholesky decomposition R of covariance matrix of error.

Robust estimate slots:

htheta:

Object of class "vectororNULL" optimized objective loss function is equal sum of rho function, with gradient and hessian as attribute.

rho:

Object of class "vectororNULL" computed robust ρ function, including gradient and hessian as attribute.

ri:

Object of class "vectororNULL" residuals equal predictor values minus predicted values, with gradient and hessian as attribute.

curvrob:

Object of class "listorNULL" robust Object of class "listorNULL" of PE and IE curvatures. Is not operational at the moment.

robform:

Object of class "nl.formorNULL", robust ρ function of object type "nl.form".

Nonlinear model estimates, inherited slots from nl.form object follows.

parameters:

Object of class "list", estimate of nonlinear model θ.

scale:

Object of class "numericorNULL", standard deviation scale estimate σ.

correlation:

Object of class "numericorNULL", correlation structure of error.

form:

Object of class "nl.form" of nonlinear model.

response:

Object of class "vectororMatrix" response, left side of formula.

predictor:

Object of class "vectororMatrix", estimated predictor η(\hat{θ)}.

curvature:

Object of class "listorNULL" of PE and IE curvatures.

history:

Object of class "matrixororNULL" convergence computations in iteration procedures, include parameters, objective function and other parameters depends on the method.

method:

Object of class "fittmethodorNULL" method of iteration used, contains main method, functions and sub methods. See fittmethod.

data:

Object of class "list" data used in computation, including response and predictor variables.

sourcefnc:

Object of class "callorNULL" source function called for fitt.

Fault:

Object of class "Fault" of error or warnings if happened.

others:

Object of class "listorNULL" of other computations, as an example the object of outlier detection measures will be saved in this slot later on.

Extends

Class "nl.fitt.rob", directly. Class "nl.fitt", by class "nl.fitt.rob", distance 2. Class "nl.fitt.roborNULL", by class "nl.fitt.rob", distance 2. Class "nl.fittorNULL", by class "nl.fitt.rob", distance 3.

Methods

parInfer

signature(object = "nl.fitt"): parameter inference function, calculate covariance matrix of parameters and their confidence interval. Usage: parInfer(object,confidence = .95)

predictionI

signature(nlfited = "nl.fitt.gn"): prediction interval. Usage: predictionI(nlfited,confidence=.95,data=NULL), data is new data that will be predicting the values for them.

residuals

signature(object = "nl.fitt.gn"): residuals of fitt.

atypicals

signature(nlfited = "nl.fitt"): detect atypical points by calculating outlier detection measures. Usage: atypicals(nlfited)

Note

All information of a generalized nonlinear robust fited model are saved in nl.fitt.rgn, thus it can be large variable of informations. It is inheritance of nl.fitt.rob, and robust form of nl.fitt.gn. It include heterogeneous or autocorrelated fitt. The heteroscedastic fitt result stores in hetro slot, and autocorelation result stores in correlation,autcor slot. meanwhile the vm, rm include contains general of covariance and correlation matrix of both heteroscedastic and autocorrelated informations.

Author(s)

Hossein Riazoshams, May 2014. Email: [e-mail] URL http://www.riazoshams.com/nlr/

References

Riazoshams H, Midi H, and Ghilagaber G, 2018,. Robust Nonlinear Regression, with Application using R, Joh Wiley and Sons.

See Also

nl.fitt, nl.fitt.gn, fittmethod, nl.fitt.rob, Fault, nl.form

Examples

showClass("nl.fitt.rgn")

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