Welcome to CSI 970, Spring 1996

Alternatives to Least Squares II

PostScript versions of the lectures, ASCII files of data and S-Plus functions, and other files are available for downloading either through Mosaic or by anonymous ftp at

We will use the book Alternatives to Least Squares, by Birkes and Dodge. Over the course of the semester each participant will give one or more lectures. The lectures will incorporate live sessions with S-Plus, Matlab, or other software.

A major objective of the seminar is to design and perform a simulation study of the performance of some tests of parameters in a linear model, using statistics that are based on various robust procedures.

The schedule is:

• January 25
  Sameena Salvucci: Results of some power comparisons
  Some handouts:
  • huber.txt M-estimator of location with MAD scale. (S function with documentation.)
  • huber2.txt M-estimator with choice of location and scale. (S function.)
  • mad.txt Robust scale estimator. (S function with documentation.)
  • power.txt Monte Carlo S program.
  • powplot.txt Monte Carlo powers for LS, L1, MH, M, and rlm fits.
  • results.pow Results of Monte Carlo study.
  • rlm.txt Robust linear regression, either Huber's M, or Huber's Proposal 2.

  ---

• February 1
  Sameena Salvucci: Further discussion of some power curves

  ---

• February 8 - 22
  Clifton Sutton: Bootstrap methods ...
  General intro to the bootstrap
  Possible tests based on the bootstrap and using the robust regression fits.