Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
4 Cards in this Set
- Front
- Back
- 3rd side (hint)
|
Accept Reject |
1. Sample U from unif(0,1) and V from pdf of v. 2. Accept if U <1/M f_Y(v)/f_V(v) where M is finite upper bound on ratio between pdf of y and pdf of v |
V must have same support as target Y and fatter tails |
|
|
Inverse CDF |
Solve for the inverse CDF of variable of interest Y. Draw from uniform and plug into the inverse CDF. |
|
|
|
Metropolis-Hastings |
Define proposal dist V. Initialize draw Z.0 from pdf of V. 1. Draw U from unif(0,1) and V.i from the pdf of V. 2. Calculate acceptance ratio (ratio of pdf of Y over pdf of V at V.i divided by ratio at Z.0) 3. Accept V.i if U is less than acceptance ratio. |
Use if ratio between target and proposal PDFs is not finite |
|
|
Importance Sampling |
-Draw theta~g(theta) and calculate weight p(theta)/g(theta) -Weight each quantity of interest h(theta) with its weight and average -Can renormalize importance weights if target p(theta) known up only to norm constant |
|
