I have been running an online Bayesian statistics course called "Bayesian Data Analysis for Climate Model Evaluation". If you would like to receive more information about the course, just send me an email.
Bayesian Data Analysis for Climate Model Evaluation
Tutor: Dr. Michel d. S. Mesquita (BCCR/Uni Research, Norway)
E-mail: michel.mesquita@uni.no
Contact info: https://sites.google.com/site/mmeclimate
Format: this course is delivered online (via a weekly email)
Week 1 – March 22: Bayesian history
A brief history of the Bayesian theory will be provided with an interesting example from Laplace.
Week 2 – March 29: Introduction to R
We will study R during this week. It would be great if everyone could install R on their computers. R can be run in any system (Windows, Linux, Mac). You can download it at r-project.org
Week 3 – April 5: Discrete variables, binomial distribution
We will revisit the binomial distribution, but this time, we will use R to make the calculations for us.
Instructions:
R file: make sure you read the instructions inside this file (open it using vim or the R text editor: File --> Open Document)!
You can run it directly in your R program in two ways:
1) by typing: source("discrete_binomial.R")
But if your file is located somewhere else in your computer: source("C:/pathtodirectory/discrete_binomial.R")
2) by opening the file using vim or the R editor and copying and pasting the information there in R
Handout file: The handout file contains the exercises for this week. Exercise 1 is solved for you in R in the discrete_binomial.R file above
Week 4 – April 12: Discrete variables, Poisson distribution
The Poisson distribution is extremely important to learn. It is used for count data. For example, when we want to model and estimate the number of hurricanes, number of storms, number of particles and other count data. We will start with the discrete form of this distribution.
Week 5 – April 19: Continuous variables, binomial distribution
Here, we will learn how to manipulate the Bayes’ rule to work with continuous variables.
Download material for this week here:
- R code for the beta distribution
- Beta distribution plot with different alpha and beta parameters (using R)
Week 6 – April 26: Continuous variables, Poisson distribution
We will expand the ideas we learned in Week 4, but this time for continuous variables. We will also make plots and estimates using R.
This week's video lecture and activities are posted on my Moodle course website at m2lab.org
Week 7 – May 3: Analyzing GCM output (Tebaldi’s method)
This is an important week. We will go through the method by Claudia Tebaldi to analyze multi-model ensembles of future climate projections against the current climate.
Week 8 – May 10: Analyzing GCM output using R
We will learn how to work with Tebaldi’s method in R.
Week 9 – May 17: Analyzing RCM output (Mesquita’s method)
Here, we will learn my method to analyze output from Regional Climate Models and to compare simulations.
Week 10 – May 24: Analyzing RCM output using R and Monte Carlo sampling
This is the last week of our course. Here, we will continue studying the method in Week 9, but this time using the Monte Carlo sampling and preparing the cloud-error plots.