Quick-Start Tutorial > Example 3: Estimating Probability with Exponential Kriging

Example 3: Estimating Probability with Exponential Kriging

Example 1 and 2 present us how to establish Kriging estimation models through setting different parameters. However, the range models from the above estimation models are still not enough in decision making. In this example of the distribution of heavy metal pollution, although we have established an estimated region model of heavy mental consistency, heavy mentals will not damage our health until the consistency reaches a certain degree. If we would like to know whether the mental consistency of each region has reach a “danger” level, we can use Exponential Kriging estimation methods to establish a model to assit decision making.

Exponential Kriging method does not establish a model directly through the values of data points, but through setting a “threshold value” to transform parameters into 0-1 mode. “0-1” is used to present whether data values are higher or lower than a threshold value. In the example of heavy mental consistency, if we assume the threshold value of the danger is “20”, the values of data points are lower than this threshold value is set as “0”, and that of data points is lower than the threshold value is set as “1”. Now we are going to establish the semivariogram model with the 0-1 binary representation of the Exponential Kriging.

1.First, open Spatial Statistical Analyst Toolbar and click Kriging to pop up Choose Input Data dialog box. Set up the “Name”, “Input Data”, “Value”, “X Field”, and “Y Field” of data source.

2.Click Next to adjust data in Adjust input data dialog box. Please select Indicator in Transform in this example.

Since data values have been transformed and presented in 0-1, the model shows us the 2 extreme values in the semivariogram model.

3.Click Next to pop up the Corss- Validation dialog box.

In this Cross-Validation dialog box, all actual values are transformed into 0-1 binary representation.

4.Click OK, and then we can obtain estimations established by the exponential Kriging estimation model.