6.4 Pros and cons

BD plots offer a model-agnostic approach that can be applied to any predictive model that returns a single number for a single observation (instance).

Advantages:

• Plots are generally easy to understand
• Illustrations are compact

Disadvantages:

• May be misleading for models including interactions (plots only show the additive attributions)
• The choice of the ordering of the explanatory variables that is used in the calculation of the variable-importance measures is important
• For a large number of variables, BD plots may be complex and include many explanatory variables with small contributions to the instance prediction.