In Evince, observations or variables can be divided into different categories. For example, an identifier that denotes the make of cars can be set as a category variable. This means all the different car makes will correspond to classes of the category "make". A category is added from the table- or plot menus.

The default setting in Evince is to create the category from identifiers or data. It is possible to constraint the names of the classes by selecting a certain part of the name that appears in the box. The user can also choose the constraints by giving the exact start position and length.

Selected objects can also be set to a certain class by choosing "Create classes from selection". In this case, the name of the class has to be given.

Right-clicking on a category in the DataSet table will bring up the table menu where the first option is a category menu for that particular category. The name of the category will be shown in italic.

Category menu options:

Set selected to...: Sets the selected observations/variables to belong to a class, to no class or to a new class.

Delete class: Removes a class.

Rename class: Renames a class.

Create dummy variable: This option is only available for observation categories. Controls if dummy variables are to be created from the category. If "one for each class" is selected, the values of each dummy variable will be either "0" or "1" depending on whether the observation belongs to the actual class. If "one for all " is selected, the observations belonging to the class will be set to "1", the second to "2", the third to "3" etc. Please note that setting the class to the second option will only create a single variable, while the first setting will create as many new variables as there are classes. In most cases, one dummy variable should be created for each class instead of creating a single dummy variable. If none is selected then no dummy variable is created from the class. Dummy variables created from a category are useful as Y variables in PLS-DA models between the X variables and the actual classes. The created dummy variables will show up in the variables view of the DataSet, shown below.

Set Equal Class Sizes: Will set equal number of observations to be included in each class. This option is particularly useful for PLS-DA models with uneven class sizes. Without this option, the PLS-DA modelling may not be successful.

Delete modification: Removes the entire category.

Rename modification: Renames the entire category.

Right- or left-clicking on the category column will also reveal category settings in the settings panel. The user can here rename or delete the category. Using the arrow button, it is also possible to select a certain class from the list and then setting selected objects to belong to the selected class. The user right-click on a class in the list and choose options from the menu. Available options are; Add new class: will add a new class to the category, Select from class(es): will select all objects from the selected classes, Rename class(es): will rename the selected classes, Change color for class(es): will option a color browser for each of the selected classes, Delete class(es): will delete the selected classes, Merge classes: will merge two or more classes into one class.

The first three buttons in the toolbar at the bottom of the list will add, rename and delete classes. The fourth and fifth arrow buttons can be used to change the order of the classes.