Evaluation method of KNN classifier

Greeting,

I am using the KNN-Classifier on Sentinel-2 images and I am getting my accuracy assessment results through the “Evaluate classifier” option. The problem is that I can not find anywhere in the documation the method that is used for the evaluation. For example, is it cross-validation, random data seperation into tarining and validation? With how many iterations and/or what percentages etc.?

Can anybody help me with that or point me to any documantion that maybe I didn’t find?

Thank you,
Spyros

You don’t mention which documentation you did find. KNNClassifierOp has a link to the Wikipedia article. The SNAP source code is the definitive reference: KNNClassifier.java (and supporting routines).

The R-project’s libraries often have links to helpful resources. The R “class” library includes a KNN classifier.