8:22 p.m.
On Wed, 3 Nov 2010, artur tarassow wrote:
Dear gretl community,
I am trying to program some dynamic multipliers on gretl.
Unfortunately, I've got some problems. Consider the following part of
the script:
-----------------------------------
scalar h = 60
scalar k = 1
matrix mtheta = {0.035316 ; -0.021316 ; -0.012700}
mtheta
mpsi = zeros(h,k)
mpsi[1,] = mtheta[1,]
mpsi[1+1,] = vphi[1:1]*mpsi[1:1,] + mtheta[1+1,]
mpsi[2+1,] = vphi[1:2]*mpsi[2:1,] + mtheta[2+1,] #Stops here
Matrix indexation ranges have to be left-to-right or
top-to-bottom, but if you need the values backwards you can
use, e.g.,
mreverse(mpsi[1:2,]) # as if mpsi[2:1,]
Allin Cottrell
12:52 p.m.
New subject: issue
Dear Sven and Allin,
Thanks for the reply. I am not sure whether I understood your comments
correctly.
Let me post some more stuff of the code:
----------------------------------------------
function matrix mplier(scalar a, matrix b, matrix varphi, matrix vpi,
scalar l, scalar h, scalar k)
vphi = zeros(1,l)
vphi[1] = 1 + a + varphi[1]
loop for (i=2; i=l-1; i+=1)
vphi[i] = varphi[i] - varphi[i-1]
endloop
vphi[l] = -varphi[l-1]
mtheta = zeros(l+1,k)
mtheta[1,] = vpi[1,]
mtheta[2,] = vpi[2,] - vpi[1,] + b
loop for (i=3; i=l; i+=1)
mtheta[i,] = vpi[i,] - vpi[i-1,]
endloop
mtheta[l+1,] = -vpi[l-1,]
mpsi = zeros(h,k)
mpsi[1,] = mtheta[1,]
loop for (i=1; i=l; i+=1)
mpsi[1+i,] = vphi[1:i]*mpsi[i:1,] + mtheta[1+i,]
endloop
mpsi[1+1,] = vphi[1:1]*mpsi[1:1,] + mtheta[1+1,] #run the loop manually
mpsi[2+1,] = vphi[1:2]*mpsi[2:1,] + mtheta[2+1,] #run the loop manually
mpsi[3+1,] = vphi[1:3]*mpsi[3:1,] + mtheta[3+1,] #run the loop manually
loop for (i=l+1; i=h-1; i+=1)
mpsi[i+1,] = vphi*mpsi[i:i-l+1,]
endloop
end function
-----------------------------------------------
The first two blocks run fine and I obtain the same results as in
gauss. The third block makes trouble. The corresponding Gauss code -
which works fine - for the third block is:
GAUSS-------------------------------
mpsi = zeros(h,k) ;
mpsi[1,.] = mtheta[1,.] ;
for i (1,l,1) ;
mpsi[i+1,.] = vphi[1:i]*mpsi[i:1,.] + mtheta[i+1,.] ;
endfor ;
for i (l+1,h-1,1) ;
mpsi[i+1,.] = vphi*mpsi[i:i-l+1,.] ;
endfor ;
----------------------------------------
There I dont obtain an error running the loop. I am really not sure
how to translate the GAUSSian code lines into proper gretl-language.
The mreverse command did not help really.
Best,
Artur
1:08 p.m.
New subject: issue
Am 04.11.2010 13:52, schrieb artur tarassow:
Dear Sven and Allin,
Thanks for the reply. I am not sure whether I understood your comments
correctly.
...
The first two blocks run fine and I obtain the same results as in
gauss. The third block makes trouble. The corresponding Gauss code -
which works fine - for the third block is:
GAUSS-------------------------------
mpsi = zeros(h,k) ;
mpsi[1,.] = mtheta[1,.] ;
for i (1,l,1) ;
mpsi[i+1,.] = vphi[1:i]*mpsi[i:1,.] + mtheta[i+1,.] ;
endfor ;
for i (l+1,h-1,1) ;
mpsi[i+1,.] = vphi*mpsi[i:i-l+1,.] ;
endfor ;
----------------------------------------
There I dont obtain an error running the loop. I am really not sure
how to translate the GAUSSian code lines into proper gretl-language.
The mreverse command did not help really.
I'm not sure what mpsi[i:1,.] does exactly in Gauss for i>1, but in
gretl mpsi[i:1,] is only allowed for i<=1 (well, and i>=0 I guess).
So depending on what Gauss really does there, you want either mpsi[1:i,]
or mreverse(mpsi[1:i,]), AFAICS. So why doesn't mreverse help?
cheers,
sven
1:21 p.m.
New subject: issue
On Thu, 4 Nov 2010, artur tarassow wrote:
Dear Sven and Allin,
Thanks for the reply. I am not sure whether I understood your comments
correctly.
Let me post some more stuff of the code:
[...[
mpsi[1+i,] = vphi[1:i]*mpsi[i:1,] + mtheta[1+i,]
The problem here seems to be that "mpsi[i:1,]" is not legal in gretl if
i>1. In gauss, "x[4:2]" is equivalent to "mreverse(x[2:4])" in
gretl.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti
10:46 a.m.
New subject: issue
Thank you for your replies Roccardo, Sven and Allin!
You were right Riccardo, the syntax in Gauss is just the other way around.
I've got another issue with the "loop for" command. I attached a part
of the code. The three loops do not seem to work for some reason. If I
run the same procedure line by line everything works fine.
I use gretl cvs (2010-10-04) on WinXP.
---------------------------------------------------------------
scalar p = 4
scalar q = 4
scalar l = maxc(p|q)
scalar h = 60
scalar k = 3
/******************************************************************************
Estimates
*******************************************************************************/
a = -0.055581
matrix b = { -0.095324 ; -0.095324 ; 0.13404}
b = b'
matrix varphi = { -0.27069 ; -0.01546 ; 0.2192 ; -0.04216 }
varphi = varphi'
matrix vpi = { -0.46627 ; 0.088921 ; 0.12036 ; 0.15333 ; -0.02255 ; -0.46627 ; \
0.088921 ; 0.12036 ; 0.15333 ; -0.022546 ; 0.70411 ; 0.23645 ;
0.22876 ; -0.8828 ; \
0.02144 }
vpi = mshape(vpi, p, k)
#Loops:
vphi = zeros(1,l)
vphi[1] = 1 + a + varphi[1]
loop for (i=2; i=l-1; i+=1) #This loop does not work
vphi[i] = varphi[i] - varphi[i-1]
endloop
vphi[l] = -varphi[l-1]
mtheta = zeros(l+1,k)
mtheta[1,] = vpi[1,]
mtheta[2,] = vpi[2,] - vpi[1,] + b
loop for (i=3; i=l; i+=1) #This loop does not work
mtheta[i,] = vpi[i,] - vpi[i-1,]
endloop
mtheta[l+1,] = -vpi[l,]
mpsi = zeros(h,k)
mpsi[1,] = mtheta[1,]
loop for (i=1; i=l; i+=1) #This loop does not work
mpsi[1+i,] = vphi[1:i]*mreverse(mpsi[1:i,]) + mtheta[i+1,]
endloop
-------------------------------------------------------------------------
Cheers,
Artur
2010/11/4 Riccardo (Jack) Lucchetti <r.lucchetti(a)univpm.it>:
On Thu, 4 Nov 2010, artur tarassow wrote:
> Dear Sven and Allin,
>
> Thanks for the reply. I am not sure whether I understood your comments
> correctly.
>
> Let me post some more stuff of the code:
[...[
>
> mpsi[1+i,] = vphi[1:i]*mpsi[i:1,] + mtheta[1+i,]
The problem here seems to be that "mpsi[i:1,]" is not legal in gretl if
i>1.
In gauss, "x[4:2]" is equivalent to "mreverse(x[2:4])" in gretl.
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
4916
days inactive
4921
days old
4 comments
4 participants
participants (4)
-
Allin Cottrell -
artur tarassow -
Riccardo (Jack) Lucchetti -
Sven Schreiber