Home Image Processing Classification Random Forests Random Forest Model Validation
See also: How to Create Classifiers, Dataset Editor, Applying an RDF Model, Variable Importance, Generating Spectral Mixtures, CreateRndForestModel, Create Binary Datasets, Encryption of a Random Forest Model 

Random Forest Model Validation

Command: Image Analysis > Classification > Random Decision Forest > Designer

The validation of random forest models is straightforward: after opening the random forest designer, click the tab "RF model validation". Next select the model to be tested and the test data to be used. Finally click the button "Validate Model" to start the validation.

As a result the tool calculates various performance measures of the selected classifier. In particular, the following information is provided:

  • several global metrics
  • several metrics for each class
  • a display of the confusion matrix per class
  • the global confusion matrix for all classes together