I have developed a vector layer which predicts which polygons are more or less likely to experience coral bleaching. I now have another vector layer with the same extent with actual coral bleaching data. Each of the layers are split into the same distribution of polygons and each polygon has a value of 1 (high), 2 (medium) or 3 (low) for both datasets.

I want to validate my predicted layer with the actual results in a quantitative manner but I'm unsure of the best way to go about this.


You could perform a spatial join between the Actual and Predicted polygon layers, to give an output layer which has fields for both Actual_Bleaching and Predicted_Bleaching.

You can then use Select By Attributes to find polygons where your predictions were correct, or where you over/under-predicted bleaching.

In terms of quantifying the result, you could calculate the number/percentage/area of the polygons which were correctly-, under- and over-predicted.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.