Apr
30
awarded  Revival
Apr
16
reviewed Reviewed How to determine predictor importance at each level of an HLM model
Apr
16
reviewed Reviewed What is the role of k-fold cross validation?
Apr
16
comment What is the role of k-fold cross validation?
Why do you think there is a disagreement here?
Apr
12
reviewed Reviewed how to Find the PROPORTION of a population for a STANDARD NORMAL between infinity and 1.25?
Apr
12
reviewed No Action Needed “Concept Class” in Machine Learning
Apr
12
reviewed Reviewed Non-algebric curve-fitting along weighted pointcloud (if possible using python)
Apr
12
reviewed No Action Needed “substituting” random variables in conditional probability?
Apr
12
revised Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
Original answer was meant to be a comment; here's the real answer.
Apr
11
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
I will do so soon. Glad we sorted that out!
Apr
11
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
I think you’re hung up on this idea that the Wold decomposition is supposed to produce a “good fit” where “good” means “high R2”. That just doesn’t follow.
Apr
11
answered Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
Apr
11
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
Or, if you prefer, it’s an $AR(\infty)$ process where the AR coefficients are all zero.
Apr
11
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
An MA(1) would include one lag of $\epsilon$. I definitely do not mean that. AR(0) fits the data here.
Apr
11
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
I agree, and that’s why I corrected myself. Your notation was unfamiliar to me and I misread it originally. Your model is essentially $ Y_t = \epsilon_t $, or “MA(0)”. It fits the data and is a Wold decomposition.
Apr
10
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
Sorry, I mean $b_0=1$.
Apr
10
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
But you have an MA model that fits: $Y_t = \eta_t$, which is what you expected.
Apr
10
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
Isn’t that what you should expect? Didn’t you in fact fit an AR(p) model, but find that all the $b_j$ are indistinguishable from zero?
Apr
10
comment Why doesn't the Wold's decomposition theorem imply a good AR(p) fit?
You haven’t shown any results. Why do you think it’s not a good fit?
Apr
9
reviewed Reviewed Deriving the Cointegrating Equation in a VECM model