Week 6 Flashcards
What is the seemingly unrelated time series equations (SUTSE) model? What is interesting about these models?
Show that the first difference of the seemingly unrelated time series equations (SUTSE) model is stationary.
When is a multivariate LLM homogeneous? What does this imply?
In a multivariate LLM what happens when Σ_η is not full rank?
What is the multivariate Common Levels model?
What is the General Multivariate (Kalman Filter) model? Why can be computing the Kalman Gain be slow?
What are the three solutions for computing the Kalman Filter for a big Multivariate model that cannot be computed because it is too slow?
Please explain what to do for a Multivariate Kalman Filter model in this case.
What is the element-by-element approach for dealing with a multivariate model?
What is the difference between updating over the series and updating over time? (Multivariate Kalman Filter model)
Please explain what to do for a Multivariate Kalman Filter model in this case. (2 options)
Please state what to do for a Multivariate Kalman Filter model in this case. (just one word)
Collapsing
What is the Dynamic Factor Model?
Write the Dynamic Factor Model in State Space form.
What transformation is done when collapsing?
What is the main idea of collapsing?
Show that when collapsing y_t* and y_t+ are independent of each other. What does this imply?
What condition holds that shows that collapsing is a valid method?
How is parameter estimation done for a collapsed multivariate model? (3 steps)
