Re: [Gretl-devel] bug which is new in 1.7.4
Am 03.04.2008 13:21, Ignacio Diaz-Emparanza schrieb:
El Thursday 03 April 2008 13:18:14 Ignacio Diaz-Emparanza escribió:
> Well, I think your backend script is very long and the hunting is
> difficult, but I have run it several times. I don't know if this may help:
>
> I have changed the value of numofW (number of bootstrap replications) and I
> see that if numofW > 20 the program hungs, but not when numofW<20. In this
Sorry, gretl hungs with numofW>=20 and not with numofW<20.
That's because of the quantile()-function and the fact that I'm asking
for the 5%- and 95%-quantile of the bootstrap distribution. (1/20 and
1-(1/20)).
Maybe the quantile()-function should be extended to also handle cases
like that? But obviously -\infty and \infty are not really useful
options, and taking the min and max of the dataset could be very
misleading for small numbers of observations, so maybe it's better to
flag an error as is currently done -- perhaps a little more informative
than just "data error".
-sven