Sam Borms

Extract feature contribution per prediction

2

Visualize contribution of features to a single prediction (or average contribution for multiple predictions)

Add profiling & profit graphs

1

Add functionality for easy plotting of profiling & profit graphs

Add at least 1 more complex benchmarking model, or facilitate integrating external models into a comparison

Raise warning when categorical variable left unprocessed

2

Add a clear warning if categorical variable is returned untouched after preprocessing (no significant categories to keep).

Make variable naming (and other) in plots customizable

1

Make names of variables, titles, axes, ... in plots maximally customizable.

When only a single category is put into Other, it makes more sense to keep the naming of that category.

Add functionality for smart testing of feature interactions (e.g. via decision trees)