Stat 740, Spring 2017

STAT 740: Statistical Computing, Fall 2017

Syllabus.
Instructor: Tim Hanson. E-mail: hansont@stat.sc.edu.
Office Hours: Tuesday/Thursday 10am to 11am, and by appointment.
Office: 219C LeConte College, (803) 777-3859.
Class Meeting: 8:30am-9:45am Tuesday/Thursday, LeConte College 201A.
Textbook: Monte Carlo Statistical Methods, Second Edition by Christian P. Robert and George Casella.
Strongly recommended: Computational Statistics, Second Edition by Geof H. Givens and Jennifer A. Hoeting.

Lecture schedule, notes, and homework

  • Thur., Aug. 24: Chapter 1 in Robert & Casella (R & C), Chapters 1 and 2 in Givens & Hoeting (G & H). Logistics, syllabus, normal approximations, Newton-Raphson. Course notes. Some examples in R.
  • Tues., Aug. 29: Newton-Raphson, multivariate delta method, examples in R.
  • Thur., Aug. 31: Large-sample Bayesian inference. Ferguson's (1973) paper on Bayesian nonparametric solutions to several problems.
  • Tues., Sep. 5: Large-sample Bayesian inference finished.
  • Thur., Sep. 7: Chapter 2 in R & C; Chapter 6 in G & H. Methods for simulating random variables. Course notes. Some examples in R. Homework 1 is due Tuesday, Sept. 19.
  • Tues., Sep. 12: CLASS CANCELLED due to Hurricane Irma.
  • Thur., Sep. 14: Methods for simulating random variables, finished.
  • Tues., Sep. 19: Chapter 3 in R & C; Chapter 5 in G & H. Numerical integration. Course notes. Some examples in R.
  • Thur., Sep. 21: Some notes on Gaussian quadrature vs. Monte Carlo integration. Homework 2 is due Thursday, Sept. 28.
  • Tues., Sep. 26: Numerical integration finished. Monte Carlo integration. Course notes. Some examples in R.
  • Thur., Sep. 28: Chapter 4 in G & H. Missing data and the expectation-maximization (EM) algorithm. Course notes from Givens and Hoeting (2013). Tim's E-M notes with R code. Original E-M paper. Laird & Ware (1982).
  • Tues., Oct. 3: E-M finished.
  • Thur., Oct. 5: Guest Professor Lianming Wang will start Markov chain Monte Carlo methods. Some R examples coded by hand.
  • Thur., Oct. 12: MCMC with examples. Homework 3 is due Tuesday, Oct. 17.
  • Tues., Oct. 17: Tim's MCMC notes with R code.
  • Thur., Oct. 19: Fall Break, no class!!!
  • Tues., Oct. 24: Some handouts from a short course: the theory behind MCMC and MCMC algorithms and WinBUGS.
  • Thur., Oct. 26: Professor Geyer's argument for one long MCMC run. The product of two normal densities is in Section 8.1.8 in The Matrix Cookbook. Here is the JAGS user manual. Homework 4 is due Thursday, Nov. 2.
  • Tues., Oct. 31: MCMC continued. Read Chapters 7 and 8 in G & H; Chapters 6 through 12 in R & C.
  • Thur., Nov. 2: MCMC continued. Slice sampling. How to do slice sampling in practice. Here's an R package that does univariate slice sampling, adaptive rejection Metropolis, and Metropolis updates.
  • Tues., Nov. 7: Ache hunting in SAS. Note that using individual univariate "blocks" mixed much better than one large block! Hypothesis testing and model choice. Tim's model selection notes with R code.
  • Thur., Nov. 9: Testing & model selection. Paper on changing α=0.05 to α=0.005.
  • Thur., Nov. 14: Testing & model selection, continued. SSVS in JAGS Example 1 and Example 2.
  • Tues., Nov. 16: Testing & model selection, finished. Paper on stably estimating LPML and WAIC. The bootstrap. Tim's bootstrap notes with R code. Chapter 9 in G & H. Homework 5 is due Thursday, Nov. 30.
  • Thur., Nov. 23: Thanksgiving, no class!!!
  • Tues., Nov. 28: Bootstrap.
  • Thur., Nov. 30: Bootstrap, finished. Tim's notes on Penalized B-splines with R code.
  • Tues., Dec. 5: B-splines, finished. A very nice set of notes and R code illustrating the bootstrap, jackkife, kernel smoothing, splines, additive models, etc. A primer on splines, knots, and penalties.