maxLik (version 1.3-4)

maxLik-package: Maximum Likelihood Estimation

Description

This is a set of functions and tools to perform Maximum Likelihood (ML) estimation. The focus of the package is on the non-linear optimization from the ML viewpoint, and it provides several convenience wrappers and tools, like BHHH algorithm and extraction of variance-covariance matrix.

Details

“maxLik” package is a set of convenience tools and wrappers to perform Maximum Likelihood (ML) analysis. It includes a) wrappers for several existing optimizers (implemented by optim); b) original optimizers, including Newton-Raphson; and c) several convenience tools to use these optimizers from the ML perspective. Examples are BHHH optimization (maxBHHH) and utilities that extract standard errors from the estimates. Other highlights include a unified interface for all included optimizers, tools to check the programmed analytic derivatives, and constrained optimization.

From the user's perspective, the central function in the package is maxLik. In the simplest form it takes two arguments: the log-likelihood function, and a vector of parameters' start values. It returns an object of class ‘maxLik’ with convenient methods such as summary, coef, and stdEr. It also supports a plethora of other arguments, for instance one can supply analytic gradient and Hessian, select the desired optimizer, and control the optimization in different ways.

One of the most useful utility functions in the package is compareDerivatives that allows one to compare the analytic and numeric derivatives for debugging the derivative code. Another useful function is condiNumber for analyzing multicollinearity problems in the estimated models.

Examples

# NOT RUN {
## estimate mean and variance of normal random vector
set.seed( 123 )
x <- rnorm(50, 1, 2 )

## log likelihood function.
## Note: 'param' is a vector
llf <- function( param ) {
   mu <- param[ 1 ]
   sigma <- param[ 2 ]
   llValue <- dnorm(x, mean=mu, sd=sigma, log=TRUE)
   return(sum(llValue))
}

## Estimate it.  Take standard normal as start values
ml <- maxLik( llf, start = c(mu=0, sigma=1) )
print(summary(ml))
## Estimates close to c(1,2) :-)

## Example how to use maxLik in your own function and allow users
## to override the default parameters
##
## 'estimate': user contructed estimation routine
## Note: it accepts both 'control' and '...'
estimate <- function(control=NULL, ...) {
   return(maxLik(llf, start=c(1,1),
                 control=c(list(iterlim=100), control),
                           # user-supplied 'control' overrides default
                           # 'iterlim=100'
                 ...))
}
m <- estimate(control=list(iterlim=1), fixed=2)
                           # user can override default 'iterlim' and
                           # supply additional parameters ('fixed')
show(maxControl(m))
                           # iterlim should be 1
print(coef(m))
                           # sigma should be 1.000

# }