SIMCA models can be used for prediction of test sets within a hyperspectral image. The user must first create a SIMCA model container as described in the multivariate data analysis section.

In order to predict a test set, the user must select a part of the image and assign the selected observations as test set observations.

The test set will be predicted automatically if a SIMCA model is created. The resulting prediction, in orange color, is available below the SIMCA model container in the data tree.

The user can now drag the orange prediction container onto the table area in order to create a prediction table. The user can also drop the container onto the plot area in order to create a Contour 2D prediction.

The prediction table shows the number of predicted pixels for each class that is part of the SIMCA model. The user can select a row in the table that belongs to a class. This will select the corresponding pixels in the Contour 2D prediction plot.