inomics
by Sven Schreiber
Hi,
I saw the call for papers for the gretl conference on Inomics today,
thanks Ignacio!
BTW, last week I learned (or perhaps re-learned, can't remember) that
Euler lived in Berlin for quite some time. Add to that that Kronecker's
grave is pretty close to the conference venue and I was thinking that
perhaps we should put together an econometrics-themed guided tour for
after the conference on Sunday...
cheers,
sven
9 years, 11 months
probit p-value option could/should replace z-stats not marg. effects
by Sven Schreiber
Hi,
as he asked me "where are the marginal effects" a colleague alerted me
to the fact that when you request p-values with probit (and probably
other such estimators as well) you lose the display of these marg.
effects. I think it may make more sense to put the p-values instead of
the z-stats, and always keep the marg. effects on display.
In principle the current behavior should be clear from the radio button
in the dialog (though I don't know whether he used the GUI or a script)
but apparently it is still easy to miss.
thanks,
sven
10 years
Function package install: Download progress window stays open
by Sven Schreiber
Hi,
with the latest Win snapshot I just installed a package from the server
(Artur's LOGIT_HETERO.gfn to be precise) and it worked ok, but the small
download progress window doesn't close itself -- I can close it
manually, though.
No big issue obviously.
thanks,
sven
10 years
matrix to list conversion
by Sven Schreiber
Hello,
this time no bug report but a feature request.
I think there is a pretty obvious feature that's missing and that's
enabling conversion from matrices to lists, to facilitate roundtripping
list - matrix - list. What I have in mind is something like this workaround:
<hansl>
function list mat2list(matrix inmat)
list lout = null # initialize
loop i=1..cols(inmat)
series stemp = inmat[,i]
list lout += stemp
endloop
return lout
end function
</hansl>
For the direction list-to-matrix we have the wonderful {} syntax, and
section 15.9 describes how to do matrix-to-series. So why not
matrix-to-list?
The thing is that currently lists and matrices in gretl are
complementary; you cannot use matrices in estimation commands for
example, and you cannot apply many calculations to lists. So I think we
need to ensure quick and easy conversion from one type to the other.
In terms of syntax, I wouldn't mind an explicit function like mat2list()
above, to be built-in. What do you think?
thanks,
sven
10 years
suddenly "unpaired if" errors popping up
by Sven Schreiber
Allin,
with yesterday's and today's snapshots I'm suddenly getting errors about
unmatched ifs that I didn't get before. I think something got broken
recently there.
The same script works ok on a Dec-11 snapshot and (testing right now)
probably on a Dec-31 snapshot, but I saw your very recent cvs log
comments about "loopsaving" and I guess that's where it's coming from.
thanks,
sven
10 years
dialectic gretl: function argument exists and doesn't exist at the same time
by Sven Schreiber
Hi,
I upgraded to the latest snapshot (Dec 31st) and encountered a strange
behavior. Unfortunately I couldn't produce a minimal example. This is in
the context of a panel dataset with a sample restriction in place that
picks out a single time period.
Then a function of mine is called, one of the arguments is a series
named "ave_inc_gr", and the corresponding argument name inside the
function is "thresh".
Now I have the following debug code inside this function, and I report
the output as comments:
# should give 0, since 'thresh' exists in the function's scope:
check = isnull(thresh)
# but actually gives 1:
print check
# but in this very next line, 'thresh' is accepted (no error):
string mys = argname(thresh)
# and gives the correct passed argument's name "ave_inc_gr":
print mys
This feels very weird, or am I blind and missing something obvious?
Oh BTW, I also noticed something else, I was using isnull() on a panel
series and it returned a series instead of an integer. Maybe you could
also look into that.
thanks,
sven
10 years
Possible issue with genseries()
by Logan Kelly
Hello,
The following script fails on gretl cvs build date 2015-01-07 running on Win 7 64 bit while the example code in help works fine.
set echo off
nulldata 50
series x = normal()
series y = normal()
list list1 = null
string SeriesName
loop i = 1.. 10
SeriesName = sprintf("test%d",i)
list1 += genseries(SeriesName, x * y)
endloop
list list1 print
it yields the following error
gretl version 1.10.0cvs
Current session: 2015-01-07 14:20
periodicity: 1, maxobs: 50
observations range: 1 to 50
Generated series x (ID 2)
Generated series y (ID 3)
Generated list list1
Data types not conformable for operation
>> list1 += genseries(sprintf("test%d",i), x * y)
Also, I cannot figure out how to reliably replicate this. But sometimes gretl terminates unexpectedly on the second time a script is run when something (I am not sure what) is wrong with a genseries() call. Sorry I cannot yet give more info on this, but I will keep experimenting.
Thanks,
Logan
10 years
pmean() with second argument and missings
by Sven Schreiber
Hi,
I stumbled over the (correctly documented) behavior of pmean() with a
second argument: "If the optional second argument is provided then
observations for which the value of mask is zero are ignored."
This means in particular that if the mask value is missing then the obs
is *not* ignored because mask is not equal to zero. So this is not
equivalent to the condition that only obs with 'mask != 0' should be
considered, because this condition would not evaluate to True if mask is
a missing value.
Is this a bug or a feature, i.e. intended or not? As I said, the
description matches the behavior, but to my personal taste it's a bit
surprising.
thanks,
sven
10 years
