I understand that there is a consensus in the literature that Halton numbers is more efficient. As Bill Greene (2003) "Simulated Likelihood Estimation of the Normal-Gamma Stochastic Frontier Function" comments  "Numerical analysts have found that a small number of Halton draws is as effective as or more so than a large number of pseudorandom draws using a random number generator. (...) as much as 10 to 1". So, it would appear, using Halton sequences of length 100 "could be like" using RNG sequences of length 1000...

That said, I read your message while I was writing the hansl script to use mrandgen()! There is always a gretl way.

Alecos Papadopoulos PhD
Affiliate Researcher
Dpt of Economics, Athens University of Economics and Business
Foundation for Economic and Industrial Research (IOBE)
web: alecospapadopoulos.wordpress.com/
ORCID:0000-0003-2441-4550

On Mon, Oct 30, 2023 at 1:24 PM Alecos Papadopoulos <papadopalex@aueb.gr> wrote:

I have a sample of size n=1789. I need to run simulated maximum
likelihood on it, generating series of pairs of independent (w,u) per
observation to compute the value of the likelihood as an simuluated average.

Instead of using the RNG of gretl, I was thinking of using Halton
numbers [...]

Alecos, please say why mrandgen() couldn't do the job for you.

Allin Cottrell
_______________________________________________
Gretl-users mailing list -- gretl-users@gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave@gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/