This question is a natural continuation of the following question: How do I construct the score process of a Markov model and verify that it is a Martingale? In this problem, we set up as follows:

Consider the following autoregressive model: $$ X_{t+1} = \alpha_0 + \beta_0 (X_t - \alpha_0) + W_{t+1}, $$ where $-1 < \beta_0 < 1$ and $W_{t+1}$ is distributed as a normal with mean zero and variance one.

The question constructs a log-likelihood process and the associated score process:

$$ s_t(\theta \mid \textbf X) = \begin{bmatrix} (1 - \beta_0) \sum_{j=1}^t W_j \\ \sum_{j=1}^t W_j (X_{j-1} - \alpha_0) \end{bmatrix}. $$

After verifying that the score process is a Martingale, a natural follow is to compute the covariance matrix for a central limit approximation for this process and find the inverse of the matrix. How would I do this? Also, how would I interpret the one-one component of the inverse?

After looking at the Wikipedia article on the Martingale central limit theorem, it seems like this may be a little difficult given that the statement of the theorem requires the difference sequence to be uniformly bounded. Am I missing something or is there a generalization that I'm missing?


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.