Hi all,
Merely as an aside, I wouldn't bet the farm on any inferences
from GPH tests if I were you, unless the data conforms to ARFIMA
having certain properties.
GPH flops badly when employed to test certain data containing inherent bias.
best regards from Richard Hudson
- Original Mess
At present im still not 100% sure how to produce 100% true forcasts using
gretl. Ive been testing using stock data from the sp500 and nasdaq stock index's
the code im using is close_spx is sp500 data close_ndx is nasdaq stock data
loop i=31..2231
smpl $i-30 $i-1
ols close_spx const time close
On Sat, 20 Feb 2010, Clive Hudson wrote:
> At present im still not 100% sure how to produce 100% true forcasts using
> gretl. Ive been testing using stock data from the sp500 and nasdaq stock
> index's
> the code im using is close_spx is sp500 data close_ndx is nasdaq stock data
>
> loop i=31