This course is an introduction to advanced methods in modern statistics, with an emphasis on applied biological data analysis. After taking this course, you will be able to invent, implement, and apply probabilistic models that describe the signal and noise inherent in real datasets. We will cover modern statistical techniques important for dealing with large, noisy datasets of many variables, including regularization, dimensionality reduction, and deconvolution. Mostly, we will work in the Bayesian hierarchical modeling framework, using Stan as interfaced through R. We will use almost no mathematical analysis (equations) for understanding and implementing these methods, relying instead on simulation and the properties of the various kinds of randomness.
The main textbook is Doing Bayesian Data Analysis : a tutorial with R, JAGS, and Stan, 2nd edition, by John Kruschke; PDF available at this link, at least on campus (see uo library entry).
We will also refer to Experimental Design and Data Analysis for Biologists, Quinn and Keough, 2002.
The Stan User’s Guide will also be often useful.
Note: both JAGS and Stan do roughly the same job, and are very similar. Although Kruschke spends more time with JAGS (for historical reasons), we will only use Stan - so, you can skip the JAGS stuff, although you should make it up with the Stan manual.
We will make constant use of R.
In class, I will use Rstudio for demonstrations, although you don’t have to use it.
To test your setup, you should be able to compile this Rmarkdown document into something that looks like this html file (but: the formatting might differ).
The main course work will be a series of reports, some short and others longer. Short reports will be less than one page of text (more, with code), are mostly based on simulated data, and are designed to demonstrate you understand and can implement an important topic from that week. These will occur in weeks without a longer report due. The three longer reports will contain more narrative, describing and analyzing a real dataset that I will provide. All reports will be turned in as self-contained Rmarkdown documents, that I should be able to run myself.
I will assign grades to each of you, independently, based on how well your course work demonstrates completion of course goals.
I will not assign grades for attendance, but class participation and group work will be important: please come. The relative weights of each category will be:
Preliminary drafts of longer reports will be handed in for feedback partway through.
I take seriously my responsibility to create inclusive learning environments. Please notify me if there are aspects of the instruction or design of this course that result in barriers to your participation! You are also encouraged to contact the Accessible Education Center in 164 Oregon Hall at 541-346-1155 or email@example.com.
The schedule is available here.