02409 Multivariate Statistics

Practical information
Lecture notes
Additional notes
Exercises
Exam
Suggested supplementary reading

Practical information

Lecture notes

Additional notes

• Collection of problems for written exercises, here.
• SAS-examples on the examples from the book, here.
• Statistical table in PS or PDF.

Solutions to problems for written exercises will be available on CampusNet as the course proceeds.

Exercises

• Lecture 1
  • SAS intro
  As stated above, today's exercise will aim at making yourself familiar with the SAS interface and the syntax of this. All information on what to do is stated in the SAS intro.
• Lecture 2
  1. SAS introduction, part 2
  2. Principal Component Analysis different datasets, here.
  3. Problem 8, E6 and E20 from the collection of written exercises.
• Lecture 3
  1. Exercise on Factor Analysis (handwritten exercise), here.
  2. SAS introduction, part 3
• Lecture 4
  1. SAS introduction, part 3 (if you didn't have the time last week)
  2. Exercises on Canonical Correlations Analysis, here.
• Lecture 5
  1. Problems 1, 4, 6 & 7 in exam from 11-12-2007 and Problem E9 from the collection of written exercises.
• Lecture 6
  1. SAS introduction, part 4 on discriminant analysis
  2. SAS exercises on discriminant analysis (if time permits).
  3. Problem E9 and E44 from the collection of written exercises.
• Lecture 7
  1. Finish SAS exercises on discriminant analysis.
  2. Problem 2 in exam from 11-12-2007.
  3. Exercise on GLM here
• Lecture 8
  1. Problem 3 in exam from 8-12-2009
  2. Test for disappearing quadratic term using SAS (this and the next exercise is a continuation of Problem 3 in the exam from 8-12-2009). On Campusnet you can find a file with the data (see Solutions -> week 8)
  3. What happens if we omit the class statement?
  4. Problem 3 in exam from 11-12-2007
• Lecture 9
  1. SAS GLM exercise on hyperspectral image data of pig lard.
  2. Problem 9 and 10 in the written exercises
• Lecture 10
  1. Problem 53 from collection of written exercises (by hand!).
  2. Problem 17 with additions.
     In this file you will also find the data that is stated in the written exercise, such that you don't need to type it yourself.
• Lecture 11
  1. Exercise on univariate regression analysis found here
  2. Problem 4 in exam from 9-12-2008.
  3. Problem 8 in exam from 10-12-2001.
  4. Finish problem 17 from last time and familiarize yourself with SAS' influence statistics.
• Lecture 12
  1. Assignment on satellite data found here.
  2. Assigment on Hotellings T² found here.
  3. Assigment on multivariate GLM found here.
• Lecture 13
  1. Exam from 2010.
     • Feel free to skip Problem 7 as it has not been covered in the lectures and likewise with question 6.1.
     • On page 233 in the current version of the lecture notes there is an error. The correct boxed formula should be this. Be aware of this when solving question 8.1.
  2. Exam from 2011.

SAS tips and tricks

A few tips on how to make SAS easier to work with can be found here.

Useful links

• PROC CANDISC explained using Iris dataset: http://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_candisc_sect019.htm
• Interpreting PROC GLM output: http://www.ats.ucla.edu/stat/sas/output/sas_glm_output.htm
• Interpreting PROC REG output: http://www.ats.ucla.edu/stat/sas/output/reg.htm
• Using the MANOVA option for PROC GLM.
• Annotated output of MANOVA: http://www.ats.ucla.edu/stat/sas/output/SAS_manova.htm.

Exam

4 hours written examination (multiple choice). The exam questions can include questions regarding the interpretation of SAS-output.

Suggested supplementary reading

• “Applied Multivariate Statistical Analysis”, 6th ed., Richard A. Johnson and Dean W. Wichern, Prentice Hall, 2007.
• “Foundations of Statistical Analyses and Applications with SAS”, Michael Falk, Frank Marohn, Bernward Tewes, Birchäuser Verlag, 2002.