The SIMCA classification is a method for classifying observations according to a created class. The classification is based upon the distance of each classified observation to a PCA model created from a certain class. A critical distance, dcrit, is calculated for each model and refers to the distance where 95% of the observations included in the model is considered to belong to the class.

A SIMCA classification can be created by right-clicking on a DataSet and by choosing "SIMCA Model".

The SIMCA model can be created from an existing class or it can be created as an empty model.

If the SIMCA model is created from a category, PCA models for each of the classes will appear below the SIMCA model container.

If an empty SIMCA model is created, PCA models must be added manually. This is done by right-clicking on the SIMCA model and by choosing "Add external model(s)". Any saved PCA models are valid for including into the SIMCA model. It is also possible to save an entire SIMCA model container and add this to an empty container.

The added models will appear below the SIMCA model container as external models. A filled green circle with a red circle inside denotes an external model.

If a test set has been created, it is possible to create a Cooman's plot from the prediction container. Drag and drop the predictions container onto the plot area in order to create this plot.

The Cooman's plot shows the distance for the test set observations to two of the models in the SIMCA classification. The critical distance to each model, dcrit, is shown for both models as red lines. The X- and Y-axis give the distances to each of the two active models. Observations predicted as no class will be shown in red in the upper right rectangle of the Cooman's plot. Observations sufficiently close to each of the two models lie to the left and below the two red lines.

In order to select a different model to be plotted in the Cooman's plot, the data range in the settings panel can be accessed.