home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #23 / NN_1992_23.iso / spool / sci / math / stat / 2085 < prev    next >
Encoding:
Text File  |  1992-10-07  |  1.7 KB  |  39 lines
  1. Newsgroups: sci.math.stat
  2. Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!swrinde!elroy.jpl.nasa.gov!lambda.msfc.nasa.gov!destroyer!caen!rob
  3. From: rob@med.umich.edu (Robert Strawderman)
  4. Subject: Re: A question or two re: splines
  5. Message-ID: <qb4-lRC@engin.umich.edu>
  6. Date: Wed, 07 Oct 92 11:59:20 EDT
  7. Organization: University of Michigan Health Sciences, Ann Arbor
  8. Distribution: umich
  9. References:  <wtfrbz1s@csv.warwick.ac.uk>
  10. Reply-To: rob@med.umich.edu (Robert Strawderman)
  11. Originator: rob@gaspra
  12. Nntp-Posting-Host: gaspra.biostat.med.umich.edu
  13. Lines: 24
  14.  
  15.  
  16. The way you suggest is the one that I am currently using - I had a feeling that the
  17. procedure I posed would not be well-behaved - you've confirmed my suspicions.
  18.  
  19. In regard to the "knot-selection" procedure you posed, there are two such references
  20. which come to mind:  Pat Smith's paper (a NASA tech report) on using B-splines in 
  21. regression, and a paper by Kooperberg and Stone which appeared recently in one
  22. of the stat journals (Comp Stat and Data Analysis, I believe).
  23.  
  24. Thanks for your reply...
  25.  
  26. -------------------------------------------------------------------------------
  27. *******************************************************************************
  28. \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
  29.  
  30. Rob Strawderman                Email:  rob@biostat.med.umich.edu
  31. Department of Biostatistics          Office:    (313) 936 - 1002
  32. University of Michigan
  33.  
  34.               "Let no man lack for the thirst of real ale...."
  35.  
  36. ///////////////////////////////////////////////////////////////////////////////
  37. *******************************************************************************
  38. -------------------------------------------------------------------------------
  39.