nl.MLE: Nonlinear MLE

In nlr: Nonlinear Regression Modelling using Robust Methods

Description

MLE estimate of a nonlinear function. with hetro variance model function, and weights.

Usage

nl.MLE(formula, data, start = getInitial(formula, data), vm = NULL, 
rm = solve(t(chol(vm))), 
control =nlr.control(derivfree = T), 
varmodel = NULL, tau = varmodel$par, ...)

Arguments

formula	nl.form object of the nonlinear function model.
data	list of data include responce and predictor.
start	list of parameter values of nonlinear model function (θ.
vm	optional covariance matrix.
rm	optional cholesky decomposition of covariance matrix.
control	list of nlr.control for controling convergence criterions. Defaul value of derivfree is "True", force function to use derivative free methods. But it can be "False" to use derivative based, has faster convergence.
varmodel	nl.fomr object of variance function model for heteroscedastic variance.
tau	list of initial values for variance model function varmodel argument.
...	extra arguments to nonlinear regression model, heteroscedastic variance function, robust loss function or its tuning constants.

Details

Calculate Maximum Likelihood estimate in several sitautions, if varmodel is given the hetroscedastic variance consider. If vm or rm is given, weighted is computing.

Value

Depending given options different fitt object will result as follow

• if vm=NULL and varmodel=NULLrepresent homogeneous and uncorrelated erro, output is nl.fitt object generated by nlsqr or nlsnm for derivative based and derivative free method respectivly given by derivfree option.

• if vm=NULL and varmodel is given represent heteroscedastic variance case, output is nl.fitt.gn generated by nl.robhetroWM, depends on using derivative free method or no.

• if vm is given represent general covariance matrix as weight, ouput is nl.fitt.gn generated by nlsqr.gn.

Note

The objective function used in nl.robhetroWM is general form of Likelihood, thus it can generate Least Square estimate using quadratic function, which can be accessed in nl.robfuncs[["least square"]] variable. It can include parameteric variance function also. Due to compatibility it is beter to be called from nlr function rather than direct call by user.

Author(s)

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

References

Riazoshams, H,. 2010. Outlier detection and robust estimation methods for nonlinear regression having autocorrelated and heteroscedastic errors. PhD thesis disertation, University Putra Malaysia.

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

See Also

nlsqr.gn, nl.robhetroWM,nl.fitt, nl.fitt.gn, nlsnm, nlsqr, nlr.control

Examples

## The function is currently defined as
"nl.MLE"

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