EViews.com

Hello!
I hope you are in good health.I would like to simulate 500 maximum likelihood estimators of an autoregressif 2 coefficients. But i don't know how to store the estimators from the following algorith:


@monte carlo maximum de vraisemblance exacte
rndseed 1 'set seed for random number generator

!mm= 500 'number of experiments


!nn = 1000 'total length of workfile

'create workfile
create u !nn

'create a vector
@vector(!mm) estm_1
@vector(!mm) estm_2

!phi1 = 0.4 'AR 1 term
!phi2 = -0.2 'AR 2 term

!k=2 'number ar terms

'create workfile
create u !nn

'declare objects to be used in simulations

matrix(!mm,!k) Cmat 'matrix to store coefficient values over simulations

'run simulations

for !sim=1 to !mm

@declaration
rndseed 123
series y=0
smpl @first+2 @last
genr y=!phi1*y(-1)+!phi2*y(-2) +nrnd

@creation dune variable pour t=1
series d1=0
smpl @first @first
d1=1
smpl @all

@creation dune variable pour t=2
series d2=0
smpl @first+1 @first+1
d2=1
smpl @all

@condition initiale
equation eq1.ls y ar(1) ar(2)
coef(1) phi1=c(1)
coef(2) phi2=c(2)
coef(3) s2=eq1.@se^2

@ la fonction log vraisemblance
logl ar2

ar2.append @logl logl2

ar2.append var=@recode(d1=1 or d2=1,s2(1)/(1-phi1(1)^2-phi2(1)^2),s2(1))

ar2.append res=@recode(d1=1 or d2=1,y,y-phi1(1)*y(-1)-phi2(1)*y(-2))

ar2.append sres=res/@sqrt(var)

ar2.append logl2=log(@dnorm(sres))-log(var)/2

@lancer maximum de vraisemblance

ar2.ml(showopts,m=1000,c=1e-5)

show ar2.output

'store coefficients estimates
@estm_1(!sim) = ar2.@coef(1)
@ estm_2(!sim) = ar2.@coef(2)
rowplace(Cmat,@transpose(ar2.ml @coefs),!sim)

next

'reset sample
smpl @all

vector cres=@cmean(cmat)

show cres
@mtos(estm_1,seri1)
@mtos(estm_2,seri2)

You appear to have lines that start with an @ symbol. That isn't valid syntax in EViews.