Three-Toed Sloth

Fall 2010 Classes: 36-757 and 36-835

I will not be teaching data mining this fall; 36-350 is being taken over this year by my friend and mentor Chris Genovese. Instead, I will be teaching 36-757 (if you'd be interested, you're already in it*), and co-teaching 36-835 with Rob Kass. Here's the announcement for the latter:

36-835 Seminar on Statistical Modeling

First meeting: Tuesday, 24 August, 1:30 pm in Porter Hall A20A (organizational)

This course will be a weekly journal club on the principles and practice of statistical modeling, organized through the careful reading and group discussion of important recent papers. Readings will be selected by the class from sources such as JASA or Annals of Applied Statistics. Discussion will emphasize the relationship between scientific questions and statistical methods. Each week students will be required to post, in an online discussion group, one cogent question or comment about the reading, and will be required to participate in the discussion. Each student will also be responsible for leading at least one class discussion. The course is intended for graduate students in Statistics or Machine Learning. Others are welcome.

If there's interest, I'll post the reading list. Our first paper will definitely be Breiman's "Statistical Modeling: The Two Cultures" (Statistical Science 16 (2001): 199--231).

Update, 26 August: handouts for 757, which may be of broader interest.

Update, 7 September: There was interest in the 835 reading list.

*: This is the first half of "advanced data analysis", a year-long project our doctoral students do on analyzing data provided by an outside investigator, under the supervision of a faculty member. ADA culminates in the student presenting their findings in written and oral form, which serves as one of their three qualifying exams. The goal is to solve genuine scientific questions, not (or not just) to use the most shiny methodological toys. If you have some real-world data which need to be analyzed, and which seem like they might benefit from the attention of a very smart statistics graduate student, please get in touch. (I promise nothing.)

Corrupting the Young; Enigmas of Chance; Self-centered