Log-Likelihood value of systemfit object

Description

This method calculates the log-likelihood value of a fitted object returned by systemfit.

Usage

1
2
   ## S3 method for class 'systemfit'
logLik( object, residCovDiag = FALSE, ... )

Arguments

object

an object of class systemfit.

residCovDiag

logical. If this argument is set to TRUE, the residual covaraince matrix that is used for calculating the log-likelihood value is assumed to be diagonal, i.e. all covariances are set to zero. This may be desirable for models estimated by OLS, 2SLS, WLS, and W2SLS.

...

currently not used.

Details

The residual covariance matrix that is used for calculating the log-likelihood value is calculated based on the actually obtained (final) residuals (not correcting for degrees of freedom). In case of systems of equations with unequal numbers of observations, the calculation of the residual covariance matrix is only based on the residuals/observations that are available in all equations.

Value

A numeric scalar (the log-likelihood value) with 2 attributes: nobs (total number of observations in all equations) and df (number of free parameters, i.e. coefficients + elements of the residual covariance matrix).

Author(s)

Arne Henningsen arne.henningsen@googlemail.com

See Also

systemfit, logLik

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
data( "Kmenta" )
eqDemand <- consump ~ price + income
eqSupply <- consump ~ price + farmPrice + trend
system <- list( demand = eqDemand, supply = eqSupply )

## perform a SUR estimation
fitsur <- systemfit( system, "SUR", data = Kmenta )

## residuals of all equations
logLik( fitsur )

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.