Problem: Cross Validation indicates differences between measured values and predicted values from Radial Basis Functions
Radial Basis Functions are interpolation techniques which are used to interpolate surfaces that pass exactly through input points.
In some cases, when validating the geostatistical layer using the Cross Validation tool, differences are identified between the measured and predicted values.
This is expected behavior. When doing Cross Validation, one data point is removed and the rest of the data points are used as the input features for Radial Basis Functions to generate an interpolation surface and to predict the removed data point. This is iterated for each data point. Since in each iteration the removed data point does not participate in the exact interpolation, the predicted value is different from the measured value.
Solution or Workaround
To verify that the interpolation surface generated by Radial Basis Functions does pass through each input point, use the GA Layer To Points tool.
- ArcMap: Performing cross-validation and validation
- ArcMap: How radial basis functions work
- ArcMap: Radial Basis Functions
- ArcMap: Creating maps using Radial Basis Functions
- ArcMap: Visualizing radial basis functions