IBM Data WH Decision Tree Node - Class Weights

Here you can assign weights to individual classes. The default is to assign a value of 1 to all classes, making them equally weighted. By specifying different numerical weights for different class labels, you instruct the algorithm to weight the training sets of particular classes accordingly.

To change a weight, double-click it in the Weight column and make the changes you want.

Value. The set of class labels, derived from the possible values of the target field.

Weight. The weighting to be assigned to a particular class. Assigning a higher weight to a class makes the model more sensitive to that class relative to the other classes.

You can use class weights in combination with instance weights. See the topic Instance weights and class weights for more information.