Stat481 - Prof. Suess

Department of Statistics and Biostatistics

California State University, East Bay

Fall 2020

I will have office hours next week as usual and by appointment.
There is no class next week.
All assignment and the Final are due by the end of the week, Friday December 11.

Questions:
Today we will discuss the idea of Bayesian Hypothesis Testing. Bayes Factors, ROPE, etc. We will also take a look at some alternative software for doing Bayesian Analysis.
R Code:
- bayestestR.Rmd
Spotlight R packages:
- bayestestR
- easystats
Spotlight Software:

Holiday: There is no class Monday November 23 - 27, the University is closed for the Thanksgiving holiday. Class we resume on Monday November 30.
Today we are going to continue our discussion of Bayesian Hierarchical Models. Draw a picture of each event in the Rose_Garden_Event_Many.zip model. We will discuss the idea of shrikage. We will discuss the difference between running rjags and runjags, the difference is the difference between single threaded programming and parallel programming.
On Monday we will discuss Homework 7.
R Project: Updated.
- Rose_Garden_Event_Many.zip
Homework: The last homework will be a reading assignment.
Quiz: The quiz has been posted. See the Homework link.
Project: The project has been posted. See the Homework link.

Holiday: There is no class on Wednesday, University Holiday.
Examples of Bayesian Hierarchal Models.
- Hierarchical Bayesian Model for Certification of a Country as "Free" from an Animal Disease
R Project:
- Rose_Garden_Event_Many.zip

Today we will discuss two other basic Bayesian models the Poison-Gamma and the Normal-Normal models. And we will begin to discuss the ideas of Hierarchical Bayesian Models.
R Project:
- Poisson_Gamma.zip
- Normal_Normal.zip
Other books:
- BHME
- Doing Bayesian Data Analysis in brms and the Tidyverse

Midterm: Due in Blackboard Friday October 23, 2020.
R Project: To run the example code, download the .zip file, uncompress it, and then click on the .Rproj file to open the R Project in RStudio.
- Rose_Garden_Event.zip
Conference: BayesiaLab See bottom of the page for previous conferences.

Today we will go over the quiz. The solution to the quiz is now available on Blackboard.
The Midterm will be made available on Wednesday. The due date will be discussed on Wedesday. See the Homework link.
I am still working on the grading of the homework and preparing solutions for the last two homeworks. I will do my best to complete the solutions for the outstanding homeworks this week and try to have grading of the homework done before the end of the week.

On Wednesday we will have a Quiz. One Bayesian Problem similar to the problems on the simplebayes.pdf handout from Homework 2, one problem to guess the (a,b) for a Beta prior, and a modeling problem like the Practice Quiz.
This week we will introduce the Markov Chain Monte Carlo (MCMC) computational algorithms used to perform Bayesian Modeling: Metropolis Algorithm and Gibbs Sampling.
We will start with a few simple examples of Monte Carlo (MC) Integration.
R Project: To run the example code, download the .zip file, uncompress it, and then click on the .Rproj file to open the R Project in RStudio.
- BernMetrop.zip
R Notebook:
- metropolis.R
Quiz: due by Friday October 9, 2020
- Quiz.zip

On Wednesday we will have a Practice Quiz. The quiz will be mostly done in class.
R Notebook:
- Practice_Quiz.Rmd
Next week on Wednesday we will have the Quiz. It will start in class and be completed at home.
This week we will be covering interative data invariance.
On Wednesday we will start with the quiz. See the Homework link.
R Project: To run the example code, download the .zip file, uncompress it, and then click on the .Rproj file to open the R Project in RStudio.
- JagsExample.zip

Homework: The solution to Homework 2 will be posted in Blackboard this week.
Homework: Homework 4 has been posted.
Today we will be discussing the mathematical solution to the Binomial-Beta problem. On Wednesday we will be discussing how to determine a Beta prior.
Discuss this week. 1. Practice Quiz next week. 2. Quiz in two weeks. 3. Take-home Midterm in three weeks.
R Project: To run the example code, download the .zip file, uncompress it, and then click on the .Rproj file to open the R Project in RStudio.
- BernABfrom.zip

Homework: The solution to Homework 1 has been posted in Blackboard. If you have not submitted a .docx or .pdf file for Homework 1 you can submit it late for consideration. No .zip files and .Rmd files will not be run, these files are not viewable within Blackboard. Please submit a .docx or .pdf file to receive credit on Homework 1.
Homework: Homework 3 has been posted.
Today we will go over the relationship between discrete priors and and continuous priors when flipping a coin.
R Project: To run the example code, download the .zip file, uncompress it, and then click on the .Rproj file to open the R Project in RStudio.
- BernGridExample.zip

Go over the handout and Tables 5.4 and 5.5 from the book.
The links to the class Jamboards have been moved to the Handouts link. (I believe I have made these visible. If you are not able to view the Jamboards, please check to make sure you are logged into your university account. If they are still not viewable, please let me know and I will check the settings again.)

This week we will start by reviewing probability notation and start working on some problems using Bayes' Theorm.
Homework: Homework 2 has been posted.
Handout: simplebayes.pdf
Handout: Simulating Bayes Rule
- bayes.r
Spotlight video: Bayes Theorem
Spotlight video: Proof of Bayes Theorem

Today we will start by downloading the programs from the Book website Doing Bayesian Data Analysis. Click Software, with programs for the book. Scroll down to the bottom of the page and download the .zip file. Unzip.
This week we will be introducing R and beginning to talk about Probability, Random Number Generation, and Probability Simulation.
Homework: Homework 1 has been updated.
Handout:
- IntroductionR.pdf
Code Examples:
- prog1.R
- prog2.R
- prog3.R
- prog4.R
- prog5.R
- prog6.R
Handouts:
Handouts:
- coin1.R
- coin2.R