English

How To: Create a cross validation graph so that the root-mean-square-standardized error is close to 1

Summary

Creating cross validation graphs show a good fit, but the root-mean-square-standardized error is not close to 1. When many data points are plotted on a graph, the graph may become crowded and interpretations become difficult.

Instructions provided are to create a graph in other spreadsheet packages, such as MS Excel, using the cross validation output created by Geostatistical Analyst.

Procedure

  1. In Geostatistical Analyst, click Save Cross Validation...
    The output will be saved as a .dbf table.
  2. Plot 'measured' versus 'predicted' values on a graph.
  3. Verify that the root mean square standardized error is calculated correctly. Do this by using the following formula:

    [O-Image] Cross validation graph
    More information about cross validation is also available in "Using Geostatistical Analyst".