Notation
Posted: Tue Nov 08, 2011 3:00 am
Hello!
I am confused a little bit. Maybe someone can shed light into the dark for me?
I want to do an MLE. For starting values I first try to "demean" the data by estimating an ARMA(p,q) with LS.
Assume for example the AR(2) coefficient is negative. Do I then have to write
Best Regards
F
I am confused a little bit. Maybe someone can shed light into the dark for me?
I want to do an MLE. For starting values I first try to "demean" the data by estimating an ARMA(p,q) with LS.
Code: Select all
'get starting values from ARMA(p,q)
equation eq_temp.ls y c ar(1) ar(2) ma(1) ma(6) ma(12) ma(15)
mu(1) = eq_temp.c(1)
phi(1) = eq_temp.c(2)
phi(2) = eq_temp.c(3)
theta(1) = eq_temp.c(4)
theta(2) = eq_temp.c(5)
theta(3) = eq_temp.c(6)
theta(4) = eq_temp.c(7)
logl spgarch_!knots
spgarch_!knots.append @logl llspg
'Residuals of ARMA(p,q)
spgarch_!knots.append res = y - mu(1) + @nan( - phi(1)*y(-1) - phi(2)*y(-2), 0)
spgarch_!knots.append resma = res - theta(1)*resma(-1) - theta(2)*resma(-6) - theta(3)*resma(-12) - theta(4)*resma(-15)
or does Eviews 7 anticipate this?spgarch_!knots.append res = y - mu(1) + @nan( - phi(1)*y(-1) + phi(2)*y(-2), 0)
Best Regards
F