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

1660GEOSTATS: sum: classes vs. thresholds for indicator kriging

Expand Messages
  • NEFIA NEFIA
    Oct 2, 2000
    • 0 Attachment
      Hello list, My original question was if anyone had opions on the use of
      using classes of a continuous variable, not thresholds for indicator
      kriging.

      Respones are as follows:
      Pierre Goovaerts, as helpful as ever, suggested:

      Benamghar and Sonnet. 1999. Performance comparison of cumulative and class
      indicators approaches for pollution risk assessment. In
      J.~G\'omez-Hern\'andez, A.~Soares, and R.~Froidevaux, editors, geoENV II -
      Geostatistics for Environmental Applications, pages 357-368. Kluwer Academic
      Publishers, Dordrecht.

      Wingle, W.L., and E.P. Poeter, 1998, Classes vs. Thresholds: A Modification
      to Traditional Indicator Simulation, Advances in Geostatistics, 1998 AAPG
      Annual Meeting, Salt Lake City, Utah, May 17-20, 1998, which is available at
      http://www.mines.edu/fs_home/wwingle/pubs/aapg_98/class.html.

      The only potential drawback with class indicators is that the experimental
      semivariograms will tend to be more erratic than for cumulative indicators
      because they will be based on a smaller proportions of non-zero data. For
      example, if you use decile thresholds to define 10 classes, each class
      indicator variable will consist of 10% of 1 and 90% of 0, while for
      cumulative indicators this ratio will be observed only for the first and
      last threshold. It's the reason why the semivariogram for a median threshold
      (50% of 1 and 50% of 0) is usually one of the best behaved.



      Simon Kelly suggested:

      Our software (Gemcom mining and geological software - www.gemcomsoftware.com
      ) performs MIK (Multiple Indicator Kriging) in the fashion you describe and
      gives an estimate and associated probability for each grade bin for each
      cell. However, the overall grade is not given by the grade with the highest
      probability. The overall grade is the sum of the mean (median in the top
      bin?) of the bin multiplied by the probability of achieving that grade.
      Another way to see the indicator probabilities is as proportions of
      different grade bins in each block. I hope I've understood you and that you
      understand me.

      Bob Sandefur suggested:

      Traditional IK (indicator kriging) will give delta classes if you subtract
      the kriged indicators one from another ie prob(x>=0.1 and
      x<0.2)=prob(x>=0.1)-prob(x>=0.2)

      I think (but don't know) that the experimental variograms of your suggested
      intervals will be harder to interpret that i traditional IK variograms (I
      suspect they will look like white noise variograms (ie flat)). Like in IK
      where they check order relations you will have to check that the sum of your
      indicators=1 (they won't exactly unless all your classes have same
      variogram.

      I never heard of it until now but its worth a try (I would assume the early
      IK people tried this and found it was harder that traditional IK) but in
      mining where IK started we're usually interest in totals above a cutoff not
      amount between cutoffs.

      Thank you all very much. A very useful webpage about indicator kriging and
      UNCERT software is : http://uncert.mines.edu/sisim/sisim.html .

      Regards,
      Andy

      _________________________________________________________________________
      Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

      Share information about yourself, create your own public profile at
      http://profiles.msn.com

      --
      *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!