Hullo
there Isobel,
In the
mining sample situation in gold several of the mining houses in SA consider that
adjusting the zeros to 50 % of the lower detection limit is the best quick
fix.(lesser of the evils !) . Obviously one must be sure that it is not a zero
meaning value being measured
Dear list
In addition to the excellent points made by Isobel, others from Anatoly
are given below. I went through the archives of ai-geostats and found
back very interesting discussions on this point (I apparently asked a
similar question in 2001, and 2003, ... I definitely have a poor
memory). As
Title: Log versus nscore transform
My apologies. The email below accidentally only
went to Gregoire only. It turns out that I haven't quite reconnected to the list
correctly.. So...
--- Original Message -
From: Michael Grant
To: Gregoire Dubois
Sent: Wednesday, August 09, 2006
(original point by Gregoire:)
>
> > Moreover, one can frequently not be "sure" about the lognormality of
>the analysed dataset, so why would one still take the risk of using
>log-normal kriging?
>
(Anatoly:)
>
>what means "not sure"??? Pearson and Kolmogorov-Smirnov tests will be
>used :-))
Mike,
I can't speak to EPA UCLs, and I'm too far removed from the
literature at this point to make a cogent argument... but I do remember
my work characterizing the hydraulic properties of artificial soils and
there was no doubt that the soil water retention curves (tension vs water
content) w
Hi Maribeth,
Sorry, I was not more clear. I was referring to
saturated hydraulic conductivities at different locations on a site (and of
course in the same hydrogeological unit). In so far as retention
curves...do you perhaps mean logK vs. water content. Very often
portions of those curves
Peter,
While the chi-square does require binning, the K-S (Kolmogorov-Smirnov)
does not. While some do bin data for a K-S, one of the reasons many
use the K-S (or the supposedly slightly better Anderson-Darling
available in Minitab for example) is that it can be (& should
generally be) done w
Regardless of how well a lognormal model represents the distribution of
the (one realization) of data,
there are still significant issues in interpreting back-transformed
kriging predictions and their back-transformed
variances. For example, because the back-transformed mean is a
function of bot
Evan Noel Cressie's book, first published in 1994, details the complete lognormal backtransform which includes the difference between the variance of the estimates and the variance of the 'true' values in the log space. My own papers of the late 1990s carried practical verifications of this tra
Gregoire,
Isobel noted that "The parametric backtransform for lognormal kriging,
for example, includes components to ensure that the backtransform
produces unbiassed estimates in the original data space."
A significant finding in Weber and Englund (Math Geol 24-4 1992) was
that this does NOT ens
Shapiro-Wilk, Shapiro-Francia, pp-plot regression
(similar to SW) are also all stock tools regarding normality tests.
At a deeper level, I can take
Gregoire's original comment:
"Moreover, one can frequently not be "sure" about the lognormality of the
analysed dataset"
to be 'one c
Hi. I am looking for a simple way to compute the kriging variance as a
function of distance between sampling sites for a given spherical and
exponential semivariogram. I know that these variances do not depend on
the data values, but on location. Thus, it seems like this would be a
common en
Yetta If you have sub-populations, the lognormal backtransform probably wouldn't work very well -- this is one place where cross validation is extremely valuable. There are many methods of 'decomposing' mixed distributions. P.D.M. Macdonald has a nice shareware program using a maximum like
Kerry:
As long the variogram model is concave upwards and increasing (i.e. no hole
effect), the kriging variance will increase as you move away from existing
sample points. So it will depend upon the model that you use and the
location of the sample points. If the fitted variogram model is steeper
14 matches
Mail list logo