Skip to content

maximum likelihood standard deviation

3 messages · Carlo Fezzi, Dimitris Rizopoulos, Ben Bolker

#
Dear R-helpers,
Anybody knows which function can I use to comupute maximum likelihood
standard errors?

Using the function "nlm" I can get the estimate of the parameters of any
likelihood that I want (for example now I am working on a jump diffusion
process) but what about the standard error?

Is there a function that I can use to calculate the second derivate of
the likelihood function respect to the vector of parameters?

Thank you so much for your help,

Carlo Fezzi
#
look at mle() (package "stats4") and optim() (i.e., "optim(..., 
hessian = TRUE)").

I hope it helps.

Best,
Dimitris

----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven

Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/16/336899
Fax: +32/16/337015
Web: http://www.med.kuleuven.ac.be/biostat/
     http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm


----- Original Message ----- 
From: "Carlo Fezzi" <fezzi at stat.unibo.it>
To: <r-help at stat.math.ethz.ch>
Sent: Thursday, June 02, 2005 1:43 PM
Subject: [R] maximum likelihood standard deviation
#
Carlo Fezzi <fezzi <at> stat.unibo.it> writes:
if you set hessian=TRUE when calling nlm it will give you
exactly what you want -- second derivative [sic] of the
likelihood w.r.t. parameters (provided of course that your
objective function in nlm() is the negative log-likelihood).
solve() on the hessian should invert the Hessian and give
you the variance-covariance matrix (then sqrt(diag()) to
get the s.d.s)