Figure 3: Animation showing network training in geographic space.

 
 
 
 
For display the hidden nodes have been equal area stretched as individual datasets.  Output nodes are displayed using a linear equal interval stretch, colours represent the same values in each output dataset.  The weights are scaled using loge transformations and then rescaled to use all available colours.  The transformations differ for the hidden and output connections.  Actual values are not shown because it is the relative values that are important.

Click on the animation to view it full size.

The file is an animated gif. It is recommended to save the image to disk and run it using an animator rather than a web browser. Doing this should enable control of the animation for speed and direction.