Transformation

When calculating semi variograms, or trend surface analysis, and cross validation, you may want to transform the data in the analysis field before it is processed. For example, you may want to use its natural logs.

You do this by choosing options from the Transformation drop-down list. You may also need to enter controlling values in one of the accompanying prompts, i.e. Additive constant or Cutoff value. These will be enabled or disabled depending on Transformation optionsyou have chosen.

By default, Transformation is set to NONE. In this case the raw form of the data values will be used in the calculation. If you need to use one of the other transformations, select it from the list.

Natural Log

Select NATURAL LOG when you need to calculate natural logs from the raw data. You can specify an Additive Constant to normalise the distribution of the natural log data.

Since the additive constant is added to all values, it has the effect of increasing the Natural log of small values relatively more than large values; in effect rotating the graph of the natural logs of the data value to the right.

Indicator

Select INDICATOR when you intend to use Indicator Kriging. You must also enter a Cutoff value.

Uniform/Rank

Use the UNIFORM/RANK option for non-parametric or distribution free statistics. Your data will is transformed so that it has a uniform distribution (where the probability of each value occurring is equal). Samples are ranked from lowest to highest. Each observation is assigned a rank. These ranks are then transformed into a percentage value against the number of samples + 1. This means a value of 75 is three-quarters of the way along the data list.

Theoretically, the semi variogram obtained from a rank transformation is topologically equivalent to the one that would be obtained from the original data values, were they not distributed in some unusual way.

Using the UNIFORM/RANK options means that no assumptions about the data distribution are made. However, this method is not considered as powerful. Try a rank transformation if all else fails.

The prompts in the Numeric Exceptions dialog enable you to control how non-numeric data in the source file is handled.