STA 293 Impossible Problems
Instructors:
Dave Higdon, 216 Old Chem, higdon@stat.duke.edu
Michael Lavine, 218 Old Chem, michael@stat.duke.edu
Meeting Times: Tues and Thurs, 10:55 - 12:10, 025 Old Chem
Overview
We will use the semester to study some problems that have a long
and unsatisfactory history. Each student, or group of 2-3 students,
will choose a topic to research during the semester. The groups
will report progress to the whole class on a rotating basis.
Possible Topics
- MCMC and convergence diagnostics
- estimating N: number of species, number of programming errors, etc.
- causal inference
- multiple comparisons
- extrapolation and prediction
- model selection
- hypothesis testing
- data = mean (smooth) + noise (rough)
- Behrens-Fisher problem
- outliers