Week 1 Flashcards
What is the signal and what is the noise? (In a linear regression for example)
What is the Local Level Model? (univariate) As what is it also known?
What is a diffuse prior density?
Show that the univariate LLM is stationary. Additionally name the two special cases of the LLM
How to simulate the LLM unconditionally? (4 steps)
What is the definition of of the signal-to-noise ratio (of univariate LLM)? How should it be interpreted?
What does the Kalman Filter compute?
What is the difference between the filtering step and the prediction step of a local level model?
What is the first step of running a Kalman Filter?
What is the second step of running a Kalman filter?
What are all the equations of computing the Kalman filter? There are 7
What is the definition of the prediction error and the state estimation? How are these related to each other?
What does Kalman smoothing imply?
How can the Kalman smoother be computed (main idea)?
What is the difference between Filtering and Smoothing?
What are the recursions (4) given by the Kalman smoother?
How are both the Filter and Smoother derived (i.e., which technique is used to get this estimation)? When do both methods have the highest amount of uncertainty?
What is the idea of weights in the Kalman Filter? How can they be implemented?
How to deal with missing observations for smoothing/filtering?
How to deal with forcasting when filtering?
How are LLM models estimated?
