Observed Fisher information matrix on a sample from the type 1 discrete Weibull distribution
Usage
varFisher(x, zero = FALSE)
Arguments
x
a vector of sample values
zero
TRUE, if the support contains $0$; FALSE otherwise
Value
a list of two matrices: the observed Fisher information matrix, and its inverse, which contains asymptotic variances and covariances of the maximum likelihood estimators of $q$ and $\beta$
x <- rdweibull(100, 2/3, 3/2)
estdweibull(x, "ML")
IF <- varFisher(x)[[2]]
diag(IF)
diag(IF)/length(x)
# asymptotic variances of the ML estimators directly estimated from the sample