REUT: Multivariate Statistics
Summer 2009
Rocks!

Instructor: Professor Barbara Bailey

Additional Dynamic Links:

• Here is Lab1
• Here is Lab2
• Here are some Lectures (Note: 2,6,7,8,9,11,14 are a good start!)
• Here is a bootstrapping practice problem and Lab3
• For plotting a bivariate normal distribution use bvnorm_plot.r
• Want to generate MVN samples? Add the library MASS and use the R mvrnorm function.
• Here is PCA example trackw.r and The American Statistician paper
• Here is some information on LDA
  Chapter 9 Old School is also helpful.
  R lda function in the MASS library is a good start.
• K-means clustering is covered in Section 10.1 of Old School. We can use the R kmeans function.
• Trees are covered in Section 9.8 of Old School. We can use the R tree function in the tree library.
• CCA (and CDA) are in Section 11.4 of Old School. We can use the R function candisc in the candisc library. Liz used CDA in her paper.
• Here are labs on Linear Models in R and ANOVA in R
• Here is

  symsqrtinv <- 
  function(x) {
      ee <- eigen(x)
      ee$vectors%*%diag(sqrt(1/ee$values))%*%t(ee$vectors)
  }
  

• R - statistical software - Get your own FREE copy of R from the R Homepage
  Here are some R links of interest:
  • FAQ
  • Try R on the web: Rweb