0
$\begingroup$

The wold decomposition theorem says that any weakly stationary process $X_t$ can be written as $$X_t=\sum_{i=1}^\infty \psi_j\epsilon_{t-j}+W_t$$ where $\epsilon_t$ is a white noise process and $W_t$ is a stationary deterministic process uncorrelated with $\epsilon_t.$

Now if I have a weakly stationary process $X_t$ with a certain mean and autocovariance function and I have another stationary process $Y_t$ in the form of the wold decomposition theorem$$Y_t=\sum_{i=1}^\infty \psi_j^{*}\epsilon_{t-j}^{*}+W_t^{*}$$

,where the process $\epsilon_{t-j}^{*}$ is white noise uncorrelated with the stationary process $W_t^{*}$, and we know that $Y_t$ has the same mean and autocovariance function as $X_t$, does this imply that we have found the wold decomposition for $X_t$ as $X_t=\sum_{i=1}^\infty \psi_j^{*}\epsilon_{t-j}^{*}+W_t^{*}$?

Edit I have a hard time pinning down the question exactly. Maybe these related questions will help:

Is the Wold decomposition unique? Can we identify a weakly stationary process by its mean and autocovariance function? If not, what if its a linear weakly stationary process?

$\endgroup$
  • $\begingroup$ Because the right hand sides of your two equations are literally identical, then $X_t$ and $Y_t$ must be the same thing (at least if mathematical notation is intended to make any sense). I suspect you mean something else, such as that the $\epsilon_{t-j}$ and $W_t$ might not actually refer to the same things in both equations, but really refer to distributions rather than random variables. Could you clarify what you mean? $\endgroup$ – whuber Apr 25 at 13:22
  • $\begingroup$ @whuber I hope my edits help to clarify the question $\endgroup$ – Joogs Apr 25 at 13:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.