LSTAR program missing values in @LOGL

[krla1]

I am asking for help because i dont have any idea why i cannot run the program. There is an information: missing values in @LOGL series at current coefficients at observation..

Thanks for help!

Karolina.

This the program:

'adjust dimension of vectors of coefficients
coef(4) beta
coef(4) alpha

coef(1) sig
coef(2) theta

smpl @all

'define variable of interest
genr y=d(mleko_sa)

smpl @first+4 @first+4
genr yf=y

logl lstar

'adapt list of parameters and initial values
lstar.append @PARAM ALPHA(1) 0.2 ALPHA(2) -0.4 ALPHA(3) -0.2 ALPHA(4) -0.2 BETA(1) 0 BETA(2) 0 BETA(3) 0 BETA(4) 0 THETA(1) 5 THETA(2) 0.5 sig(1) 0.1

lstar.append @logl lstar1

'adapt model equation
lstar.append YF=alpha(1) +ALphA(2)*Y(-1)+alpha(3)*Y(-2)+alpha(4)*y(-3)+(BeTA(1)+beta(2)*Y(-1)+beta(3)*y(-2)+beta(4)*y(-3))/(1+EXP(-THeTA(1)*(Y(-2)-THeTA(2))))


lstar.append RES=Y-YF
lstar.append VAR=SIG(1)^2
lstar.append Z=RES/@SQRT(VAR)
lstar.append LsTAR1=LOG(@DNORM(Z))-LOG(VAR)/2


smpl @first+4 @last
lstar.ml(showopts,m=1000,c=1e-5)
show lstar.output

smpl @all

[EViews Gareth]

It means that the logl went into a region where NAs were generated, thus the estimation stopped. You could try adjusting your starting values.

[krla1]

Thank You! But tell me how can I adjust them?

Sorry I am the beginner:)

Karolina.