Statistics 4M03/6M03

Multivariate Analysis

Dr P.D.M. Macdonald

2002-2003

Calendar Description

Multivariate distributions: Normal, Wishart, T² and others; regression, correlation, factor analysis, general linear hypothesis.

• Course Outline

Announcements

Your Final Examination will be held on Monday 9 December 2002, from 9:00 am to 1:00 pm, in the BSB-244 computer lab. Exam hint: Learn the complete gamma function!

I have added an example showing how to draw an elliptical confidence region; it is on the page "Drawing an Ellipse in Splus or R".

Special Exam Office Hours: Sat 7 Dec 10:30-12:00, 14:00-16:00; Sun 8 Dec 20:00-22:00. If you can't get into the building at any of these times, telephone me at 905 525 9140 x 23423.

Timetable

Monday at 12:30, Tuesday at 13:30, Thursday at 12:30 in JHE-210.

Computer Labs

• The class will meet in the BSB-241 Computer Lab at the following times:
  • 2002-09-19 12:30 (Thursday)
  • 2002-09-23 12:30 (Monday)
  • 2002-09-24 13:30 (Tuesday)
  • 2002-09-26 12:30 (Thursday)

  You must have a valid userid, password and laser printing account for the lab!

Computer Lab Exercises

These exercises are for self-directed learning. They will not be graded and you do not have to hand them in. However, you will find that writing down your interpretation of the results is an excellent way to clarify your understanding. If you complete your report and submit it to me, I will read it and return it with comments and suggestions.

• Getting started in Splus and R
• Using Splus to draw a 3-D frequency polygon for bivariate data
• Get the datasets for Srivastava, Methods of Multivariate Statistics
• Explore the crude oil data (Excel file)

Text

• Srivastava, Muni. Methods of Multivariate Statistics, Wiley.

Muni Srivastava was awarded the Gold Medal of the Statistical Society of Canada this year. Click here to see photos of him taken in the McMaster University Student Centre.

Class Notes

• Drawing an Ellipse in Splus or R
• Plotting the Bivariate Standard Normal Density in Splus or R

Computing Resources

• R for Beginners - a 31-page introduction to R, in pdf format - and other online manuals
• Download R - a freeware package with the look and feel of Splus
  
  

  You want the "Precompiled binary distribution", not the "source code". For Windows installation, go to the link for "Windows (95 and later)", open folder "base", and click on SetupR.exe to download the installer SetupR.exe. Then run SetupR.exe to install R on your computer.

  
  
• An On-Line Splus Tutorial from Dalhousie University
  
  

  Because you are running R in Windows, not UNIX, you should skip the Introduction, which is mostly about UNIX, and begin with Entering Data.

  
  
• Insightful Corporation- Vendor of Splus
• An On-Line Introduction to UNIX
• Running UNIX versions of Splus, R, and SAS at McMaster

Assignments

• Assignment #1 - Due 2002-10-21 17:00
• Assignment #2 - Due 2002-11-25 17:00
  
  • Use the "track records" data  from Test #1. Try all the methods of Chapter 3 to test for multivariate normality and test for outliers. Try transformations to improve normality or make the data more interpretable, in particular, 1/x, -log(x), sqrt(x), Box-Cox.
• Assignment #3 - Due 2002-12-02 17:00
  • Handed out in class 2002-11-12

Tests

Some Other Links

• Statistics Seminars at McMaster
• Statistical Society of Canada / Société statistique du canada
• SORA (Southern Ontario Regional Association of the SSC)
• My work on mixture distributions

Need more help?

• If you have concerns about preparing for tests and examinations, improving your study habits, giving class presentations, or mastering English as a second language, the Centre for Student Development can help.