eq.s options in the rolling regression
Posted: Mon Dec 07, 2015 6:15 am
Hi,
I have to check the robustness of my estimates with different time period.
I am using the below code for the same:
if @isobject("comparison") then
delete comparison
endif
'create vector to store r-squares
matrix(6,64) comparison
'create empty equation to be used inside the loop
equation eq
%smpl_start="2000Q1"
%smpl_end= "2015Q2"
smpl {%smpl_start} {%smpl_end} '@last
'get size
!length = @obssmpl
!j=1
for !i = 0 to !length step 1
!j=!j+1
'smpl 1996q1+!i 1996q1+8+!i
smpl {%smpl_start} {%smpl_start}+!i+20
eq.ls 'your model equation
'store coefficients into matrix
colplace(comparison, eq.@coefs, !j)
next
show comparison
But I am not able to match my model estimates with the results with this code.
I search and found that to get the exact estimates I have to feed the same option which i select when i ran the final model.
Could anyone please help me how to specify eq.ls (options) in the above code.
My estimation settings option is - LS - least squares(NLS and ARMA)
Co-efficient co-variance : Method - Ordinary and information matrix - OPG
d.f adjustment has been ticked
Optimization method - Eviews legacy
Maximum Iterations - 500
Convergence tolerance - 0.0001
ARMA method - CLS
Starting ARMA coefficients values: OLS/TSLS
Backtest MA terms has been tikced.
Below is the Eviews help on the above topic:
http://www.eviews.com/help/helpintro.ht ... 0.068.html
Regards,
Nawal Kashyap
I have to check the robustness of my estimates with different time period.
I am using the below code for the same:
if @isobject("comparison") then
delete comparison
endif
'create vector to store r-squares
matrix(6,64) comparison
'create empty equation to be used inside the loop
equation eq
%smpl_start="2000Q1"
%smpl_end= "2015Q2"
smpl {%smpl_start} {%smpl_end} '@last
'get size
!length = @obssmpl
!j=1
for !i = 0 to !length step 1
!j=!j+1
'smpl 1996q1+!i 1996q1+8+!i
smpl {%smpl_start} {%smpl_start}+!i+20
eq.ls 'your model equation
'store coefficients into matrix
colplace(comparison, eq.@coefs, !j)
next
show comparison
But I am not able to match my model estimates with the results with this code.
I search and found that to get the exact estimates I have to feed the same option which i select when i ran the final model.
Could anyone please help me how to specify eq.ls (options) in the above code.
My estimation settings option is - LS - least squares(NLS and ARMA)
Co-efficient co-variance : Method - Ordinary and information matrix - OPG
d.f adjustment has been ticked
Optimization method - Eviews legacy
Maximum Iterations - 500
Convergence tolerance - 0.0001
ARMA method - CLS
Starting ARMA coefficients values: OLS/TSLS
Backtest MA terms has been tikced.
Below is the Eviews help on the above topic:
http://www.eviews.com/help/helpintro.ht ... 0.068.html
Regards,
Nawal Kashyap