0
$\begingroup$

I've come across a question where I have an MA(1) process like so:

$X_t = b_t - 0.4 b_{t-1}$ (where $b_t$ is a white noise process and $t$ is the time index)

The question asks me to find $\phi_{11}$ (Phi) and $\phi_{22}$

From what my notes say:

  • $\phi$ is from the AR(1) process.
  • There exists a technique to convert an MA(1) process into an infinite AR process

However, I cannot find any information or techniques for converting an MA(1) process into an AR(1) process or even for using the $\theta$ (Theta) value as a means for converting into the $\phi$ value.

Is there something I am missing in this time series problem?

For some extra info:

  • I have calculated the mean of the MA(1) process
  • I also have the variance
  • I also have autocorrelations for $k=0$, $k=1$

Any help will be much appreciated.

$\endgroup$

1 Answer 1

1
$\begingroup$

Ok, it seems I have come across something called the PACF of the MA(1) process.

I think I may have confused the Phi in the AR(1) process with this other Phi. I'm not even sure they are related.

The formulae are:

Phi11 = Rho1 = -Theta1 / 1 + Theta1^2

Phi22 = - (Rho1^2 / 1 - Rho1^2)

$\endgroup$
1
  • $\begingroup$ $\phi_i$'s and $\phi_{ii}$'s are related. $\endgroup$
    – Glen_b
    Commented Mar 7, 2014 at 0:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.