This course is the third semester of the required four semester applied statistics sequence in the School of Statistics Ph.D. program. To enroll in this course you must be a second-year Ph.D. student in statistics or have permission of the instructor.

The course meets 2:30-3:20 MWF (09/07/2012 - 12/14/2012) 206 Vincent Hall.

The handouts shown below are from Fall 2011. The syllabus will change somewhat. In particular, the first unit on mixed models will be dropped (this is part of Stat 8052 and was included for review). An additional topic or two will be added, probably on incomplete data and possibly on Tobit models. The nearly weekly assignment model used in the past will also change to approximately 6 assignments. All students will be expected to make one or two presentations to the class, depending on the enrollment.

2011 Handouts

• Sept. 7, 2011 2011 Course syllabus.
• Sept. 7 linear mixed model. R commands: fishgrowth.R.
• Sept. 12 generalized linear mixed model. R commands: glmm2.R.
• Sept. 16 smoothing with one predictor. R commands: smoothing.R.
• Sept. 23 Effects plots. R commands: effects.R.
• Sept. 28 additive models. R commands: gam.R.
• Oct. 3 generalized additive models. R commands: gam2.R.
• Oct. 5 Robust (M and L1) fitting. R commands: robust.R. quantreg vignette: rq.pdf, quantreg manual quantreg.pdf.
• Oct. 10 Recursive partitioning: rpart.R Here is a link (U of M only) to a book by R. Berk that has extended presentations of these topics: Statistical Learning from a Regression Perspective.
• Oct. 21 Artificial Neural Networks. R commands: nnet.R.
• Oct. 24 Variable Selection and Regularization. R commands: lasso.R.
• Nov. 2 Manova. R commands: manova.R.
• Nov. 7 Chapter 8, low-dimensional approximations. R commands: chap8.R.
• Rev. Nov. 17 Chapter 9, PCA. R commands: chap9.R.
• Rev. Nov. 18 Chapter 10, Factor analysis. R commands: chap10.R.
• Nov. 21 Chapter 11, Clustering. R commands: chap11-clustering.R
• Dec. 2 Chapter 12, LDA. R commands: chap12-lda.R.
• Nov. 29 Chapter 14, Canonical Correlations and seemingly unrelated regression. R commands: chap14-cancor.R.