Loading ...
Sorry, an error occurred while loading the content.

GEOSTATS: Summary of Boostrapping Replies

Expand Messages
  • William C. Thayer
    Hello. A summary of the replies I received to the following request is provided below. I have decided to go with Mathematica after discovering the huge
    Message 1 of 1 , Apr 24, 1998
    • 0 Attachment
      Hello. A summary of the replies I received to the following request is
      provided below. I have decided to go with Mathematica after
      discovering the huge discount for students. It is capable of generating
      voronoi diagrams as well as performing the resampling (and many other
      things of course).
      I have not looked into the reply regarding placing constraints on
      resampling (reply #11) although intend to. I am curious if anyone else
      has feedback on this subject.
      Thanks to all those who responded.

      ORIGINAL REQUEST:
      I am working on a m.s. thesis that involves the use of
      parametric and non-parametric methods of estimating average
      concentrations of lead. My questions are:
      1) Does anybody have or know of software (preferably shareware) that
      performs bootstrapping (or some other resampling method)? I am aware of

      Mathematica but it is not within my budget.
      2) I am also looking for software that can calculate theissen polygons
      (aka: voronoi diagrams and polygons of influence).
      Any help will be appreciated,
      Bill Thayer

      SUMMARY:
      1) Check out the " Software for bootstrap computations" appendix in "An
      Introduction to the Bootstrap" Monograph on Statistics and Applied
      Probability 57; Efron, Bradley and Robert J. Tibshirani. Chapman & Hall,
      1993. It mentions an MS-DOS package called "Resampling stats. Obtained
      from: Resampling Stats
      612 N. Jackson St. Arlington, VA 22201

      2) S or S-plus has built-in facilities for vectors, matrices, etc. S
      language functions are given in the above appendix.

      3) Are you aware of the cross-validation capabilities of kt3d in the
      GSLIB package? It will reestimate (kriging) each sample point in the
      data set using all the other points. You might also be interested in
      the paper:
      Direct Assessment of Local Accuracy and Precision, by Clayton Deutsch
      presented at the 1996 Geostats Congress in Wollongong.

      4) I just bought Mathematica for a decent price. As a student, you can
      get Mathematica 3.0 (the latest version) for $ 139.95. The phone number
      is 1-800-441-MATH. Although you can order it by phone, you will be
      asked to mail or fax a copy of your student ID. If I remember
      correctly, this price was not on the Wolfram/Mathematica Home-page, but
      it was on a brochure that I received. The only disadvantage of the
      student version is that you do not get the manuals in hard copy, but on
      CD-ROM.

      5) RESAMPLE is on the SU clusters server.

      6) I would recommend Matlab. It too may not be within your budget, but
      I would check student prices and availability at your school. There is
      a lot of code available and it is a superlative number cruncher. As a
      non-mathematician and non-programmer, I have found it very intuitive and
      easy to modify to my needs.

      7) If you have access to your university software then there is a good
      chance that they have IMSL. IMSL is a collection of Fortran subroutines
      for problem solving and statistical development.

      8) If you get a general purpose program like IDL you can easily write a
      simple program to do resampling and many other things. I think that
      they have a student version available.

      9) Techbase software has an academic loan program with minimal costs
      involved to set it up

      10) You could write your own program. "Numerical Recipes" by Press et
      al. is a good source of random number generators.

      11) Bootstrap has to be constrained by some statistical properties as
      revealed from the data available (variogram, pdf, ....etc). To do so you
      have to find a way by which you can generate a random field that
      reproduce similar statistics as estimated from the data. See: Lall and
      Sharma, WRR 32(3), 679-690, 1996.





      --
      *To post a message to the list, send it to ai-geostats@....
      *As a general service to list users, please remember to post a summary
      of any useful responses to your questions.
      *To unsubscribe, send email to majordomo@... with no subject and
      "unsubscribe ai-geostats" in the message body.
      DO NOT SEND Subscribe/Unsubscribe requests to the list!
    Your message has been successfully submitted and would be delivered to recipients shortly.