28 lines
No EOL
1,011 B
R
28 lines
No EOL
1,011 B
R
# Wald test for estimating equations
|
|
#
|
|
# Author: Francois Pelletier
|
|
#
|
|
# LGPL-3.0
|
|
###############################################################################
|
|
|
|
|
|
#' Wald test for estimating equations
|
|
#' @param param Estimated parameters
|
|
#' @param n Size of estimation sample
|
|
#' @param R Matrix of linear coefficients of the constraints
|
|
#' @param r vector of linear constants of the constraints
|
|
#' @param eqn.covariance Covariance matrix of the estimating equations
|
|
#' @param eqn.gradient Gradient matrix of the estimating equations
|
|
#' @param alpha level of confidence
|
|
#' @return A list containing the statistic, p-value and reject of the null hypothesis
|
|
#'
|
|
#' @author François Pelletier
|
|
Wald.Test <- function(param,n,R,r,eqn.covariance,eqn.gradient,alpha=0.05)
|
|
{
|
|
wald.stat <- n * t(R %*% param - r) %*%
|
|
ginv(R %*% eqn.gradient %*% eqn.covariance %*%
|
|
t(eqn.gradient) %*% t(R)) %*%
|
|
(R %*% param - r)
|
|
PV <- pchisq(wald.stat,nrow(R))
|
|
list(wald.stat=wald.stat, p.value=PV, reject=PV>1-alpha)
|
|
} |