Day 5 (January 31)
Announcements
- Please read (and re-read) Ch. 4 in BBM2L
- You may want to take a look at Ch. 6
- We will skip Ch. 5
- Assignment 2
- Review Eq. 3.1 in BBM2L
- General MC integration using uniform dist
- Don’t go to extreme lengths to make the programing more difficult
- With MC integration, the most important skill is to know what integral you are approximating
- I will not be posting a guide to Assignment 2
- Final project is posted
- Proposals are due Feb. 17
- Assignment 3 is due Sunday Feb. 5
Introduction to Metropolis-Hastings algorithm
- Why use a Metropolis-Hastings algorithm?
- Original work (see link and link)
- Only need to know a function that is proportional to the PDF/PMF
- Why this is such a big deal for Bayesian statistics?
- What else do we need to unlock the power of Bayesian?
- What we loose by using a Metropolis-Hastings algorithm
- Requires a bit more programming and supervision/checking
- Correlated samples vs. independent samples
- Burn-in interval
- Live illustration using example from homework 2 and 3