Too many residual series outputs

mattyb · Postby **mattyb** » Fri Sep 23, 2011 7:32 pm

Hi all,

First off would like to say that these forums have been very helpful, I have been following the intro into eviews programming and learning alot. I am having a problem where when i estimate my equations I am getting ALOT of residual series outputs, were I only want outputs for each of my equations in my system (ie for estimate AA_R = C(1) + .....and AA_CDS C(1) + ..... i want a residual for each (i should be getting 281x2 = 562 residual series outputs) . My coding is letting me down i know this, so any corrections would be an advantage.

Code: Select all


system zzzz

for !i=1 to 281
   %iname = zzdcds.@seriesname(!i)
 	%jname = zzr.@seriesname(!i)
%kname = dcds_1.@seriesname(!i)
%lname = zzdcds_2.@seriesname(!i)
%mname = zzdcds_3.@seriesname(!i)
%nname = zzdcds_4.@seriesname(!i)
%oname = zzdcds_5.@seriesname(!i)
%pname = zzr_1.@seriesname(!i)
%qname = zzr_2.@seriesname(!i)
%rname = zzr_3.@seriesname(!i)
%sname = zzr_4.@seriesname(!i)
%tname = zzr_5.@seriesname(!i)


zzzz.append {%iname} = C(1) + C(2)*{%iname}(-1) + C(3)*{%iname}(-2) + C(4)*{%iname}(-3) + C(5)*{%iname}(-4) + C(6)*{%iname}(-5) + C(7)*{%jname}(-1) + C(8)*{%jname}(-2) + C(9)*{%jname}(-3) + C(10)*{%jname}(-4) + C(11)*{%jname}(-5) + C(13)*{%pname} + C(14)*{%qname} + C(15)*{%rname} + C(16)*{%sname} + C(17)*{%tname}

zzzz.append {%jname} = C(12) + C(2)*{%iname}(-1) + C(3)*{%iname}(-2) + C(4)*{%iname}(-3) + C(5)*{%iname}(-4) + C(6)*{%iname}(-5) + C(7)*{%jname}(-1) + C(8)*{%jname}(-2) + C(9)*{%jname}(-3) + C(10)*{%jname}(-4) + C(11)*{%jname}(-5) + C(18)*{%kname} + C(19)*{%lname} + C(20)*{%mname} + C(21)*{%nname} + C(22)*{%oname}

zzzz.sur
zzzz.makeresid resid_{%iname}
next

Bonus points: after I run this I want to break it up into rolling windows of say 100 observations, any suggestions or links on how to do this. :)

Much appreciated!

mattyb · Postby **mattyb** » Sun Sep 25, 2011 9:24 pm

solved my own problem,

Code: Select all


smpl 1 775

for !i=1 to 281
   %iname = zzdcds.@seriesname(!i)
   %jname = zzr.@seriesname(!i)
%kname = dcds_1.@seriesname(!i)
%lname = zzdcds_2.@seriesname(!i)
%mname = zzdcds_3.@seriesname(!i)
%nname = zzdcds_4.@seriesname(!i)
%oname = zzdcds_5.@seriesname(!i)
%pname = zzr_1.@seriesname(!i)
%qname = zzr_2.@seriesname(!i)
%rname = zzr_3.@seriesname(!i)
%sname = zzr_4.@seriesname(!i)
%tname = zzr_5.@seriesname(!i)

system zzzz_{%iname}

zzzz_{%iname}.append {%iname} = C(1) + C(2)*{%iname}(-1) + C(3)*{%iname}(-2) + C(4)*{%iname}(-3) + C(5)*{%iname}(-4) + C(6)*{%iname}(-5) + C(7)*{%jname}(-1) + C(8)*{%jname}(-2) + C(9)*{%jname}(-3) + C(10)*{%jname}(-4) + C(11)*{%jname}(-5) + C(13)*{%pname} + C(14)*{%qname} + C(15)*{%rname} + C(16)*{%sname} + C(17)*{%tname}

zzzz_{%iname}.append {%jname} = C(12) + C(2)*{%iname}(-1) + C(3)*{%iname}(-2) + C(4)*{%iname}(-3) + C(5)*{%iname}(-4) + C(6)*{%iname}(-5) + C(7)*{%jname}(-1) + C(8)*{%jname}(-2) + C(9)*{%jname}(-3) + C(10)*{%jname}(-4) + C(11)*{%jname}(-5) + C(18)*{%kname} + C(19)*{%lname} + C(20)*{%mname} + C(21)*{%nname} + C(22)*{%oname}



zzzz_{%iname}.sur

zzzz_{%iname}.makeresid(n=arg{!i}) 


next

'create matrix to store cause statistics and p-values
matrix(281,4) causes

''counter of how many equations we have run
!rowcounter=1

'run pairwise regressions between each series
for !i=1 to 281
   
      freeze(causetable) arg{!i}.cause(5)        ' perform  test and freeze it into a table called causeTable
      causes(!rowcounter,1) = @val(causetable(8,3))   ' store cause statistic (which is in cell 6,2 of the table)
     causes(!rowcounter,2) = @val(causetable(8,4))
  causes(!rowcounter,3)  = @val(causetable(9,3))
  causes(!rowcounter,4)  = @val(causetable(9,4))   ' store causep-value (which is in cell 6,4 of the table)
      !rowcounter=!rowcounter+1         ' increment row counter
      d causetable    ' delete the table
next

Im upto running a rolling sample period

EViews.com

Too many residual series outputs

Too many residual series outputs

Re: Too many residual series outputs

Who is online