AI RESEARCH
Woodelf++: A Fast and Unified Partial Dependence Plot Algorithm for Decision Tree Ensembles
arXiv CS.LG
•
ArXi:2605.14578v1 Announce Type: new Partial Dependence Plots (PDPs) visualize how changes in a single feature affect the average model prediction. They are widely used in practice to interpret decision tree ensembles and other machine learning models. Joint-PDPs extend this idea to pairs of features, revealing their combined effect. Partial Dependence Interaction Values (PDIVs) measure feature interactions. The Any-Order-PDIVs task computes these interactions for every feature subset across all rows of the dataset. Woodelf++ is implemented in pure Python and s GPU acceleration.