Loading ...
Sorry, an error occurred while loading the content.
Advanced Search
Author
Subject
Message
Special notice only

15 results from messages in ai-geostats

Advanced Search
  • For estimation/simulation with a finite search neighborhood, assuming an intrinsic hypothesis becomes quite tangential to the main point of interest -- i.e., generating good estimates. Assuming something is fBm is just that -- an assumption. It might even be a valid one where variance increases without bounds, but still other such concerns as the size of the search radius, number...
    Syed Abdul Rahman Nov 30, 1999
  • There is also the well trodden path of combinatorial optimization schemes (annealing, GAs) whereby one can flexibly model the structure of variates and co-variates without having to worry about LMC. At the cost of CPU time, of course. Syed Pierre Goovaerts on 06/03/99 01:03:41 AM To: gerhard cc: ai-geostats@^$1 Subject: Re: GEOSTATS: Linear Model of Coregionalization Hi Gerhard...
    Syed Abdul Rahman/SINGPROD1/Landmark Jun 2, 1999
  • To this I am tempted to quote Cressie, "the role of cross-validation is to prevent blunders and to highlight potentially troublesome data. It cannot prove that the fitted model is correct, merely that is is not grossly incorrect." Even more important is the "reasonableness" of the eventual spatial surface, taking into account related "soft" information, and how credible the results...
    Syed Abdul Rahman Dec 6, 1998
  • Fetching Sponsored Content...
  • Now that I think about it I am not sure that it would be practical to construct a rose diagram for each pred. location, experimenting with different angular and lag tolerances and directions in order to derive some sort of anisotropy ratio, and then fit the variogram models automatically, and then make a prediction, move on to a different search neighborhood and repeat the whole...
    Syed Abdul Rahman Sep 30, 1998
  • A popular taxonomy is that of Noel Cressie's (in _Statistics for Spatial Data_, Wiley Interscience) in which he gives three separate treatments for: (a) geostatistical data (data with continuous spatial index); (b) lattice data (data with a countable collection of spatial sites, frequently regular); and (c) point pattern data (data with random spatial index). All of this under the...
    Syed Abdul Rahman Sep 29, 1998
  • There have been some attempts at "adaptive" interpolation, e.g. refer papers by Wong et al at the University of New South Wales on the use of RBF networks for spatial prediction. Initial attempts have been at optimizing RBF constants (analogous to the "range") and type of RBF (analogous to the variogram type) which differ from neighborhood to neighborhood. Anisotropy ratios can...
    Syed Abdul Rahman Sep 29, 1998
  • From: John Sadler >I have scrounged several sources of routines available on the web, but most >of those are far more sophisticated than my needs. Sophistication is relative, consider that: (a) a pure nugget model decomposes into a local mean interpolation technique (b) some specific power law variogram parameters can decompose into a linear interpolation between points. (c) there...
    Syed Abdul Rahman Jun 12, 1998
  • Some comments on Tom Charnock's queries: A) Generally, the rule of thumb is to limit the search neighborhood, to avoid stretching the stationarity assumption. Of course, this is only fine and dandy if you have sufficient data. This could range from three to six wells in an offshore development to hundreds of thousands of points in a remote sensing problem. B) Consider that cross...
    Syed Abdul Rahman Jun 12, 1998
  • Pierre Goovaerts wrote: > I would recommend to either avoid lognormal kriging, or use > stochastic simulation. The basic idea of simulation consists > of generating realizations of the spatial distribution of the target > attribute which reproduce the pattern of spatial variability of point > measurements. Block simulated values are then computed using linear or > non-linear...
    Syed Abdul Rahman May 9, 1998
  • One has to reason why logarithmic transforms are used, a common reason is to work with a manageable range of values, and another is to get a stable variogram. Nevertheless, krigings become extremely sensitive to changes in sill values, once the results are transformed to their raw values. Another problem is that such results are biased, what you expect to be the expectation...
    Syed Abdul Rahman Feb 5, 1998