9 Day 9 (February 14)
9.1 Announcements
My zoom office hours will be 2-3pm today (not the usual 1-2 pm).
See Tips and tricks (pgs. 541) and Glossary (pg. 543)
Assignment 5 is posted and due Sunday Sunday, February 19.
Class project proposal is due Friday
Assignment #4
- Metropolis-Hastings ratio
- Don’t forget about the support!
- What does
rnorm(1,2,3)
do? (i.e., \(\sigma^2\) vs. \(\sigma\)) - Question 7 and algorithm efficiency
- Optimal proposal (work out on whiteboard)
- Requires no tuning
- Metropolis-Hastings ratio
9.2 Our second statistical model
- Dig into the rabies test a bit more….
- What statistical model should I use if I know/estimate if I have rabies?
- Remember that with our first statistical model we were trying to estimate if probability a randomly sampled bat has rabies.
- Building a statistical model using a hierarchical Bayesian approach
- Specify (write out) the data model (you may want to take a look at Ch. 10 and Ch. 23)
- Specify the process model
- Specify the parameter model (or prior) including hyper-parameters
- Select and approach to obtain the posterior distribution
- Gibbs sampler (re-read parts of Ch. 4 about Gibbs sampler; pgs.35 - 37)
- Derive full conditionals
- Discussion of trade-offs with Gibbs sampler with analytical full conditionals vs. Metropolis-Hastings
- Live example (download)