Tree Ensemble Explainability

Alexander Moore, Yaxiong Cai, Kristine Jones, Vanessa Murdock

Abstract: Complex machine learned models play an increasingly important role in modern technologies, consuming large amounts of data to provide a plethora of useful services. While these systems are highly effective, many of them are black boxes and give no insight into how they make the choices they make. Moreover, those that do often do so at the model-level rather than the instance-level. In this work, we present a method for deriving instance-level explanations for tree ensemble models and examine its applications. Tree ensemble models such as Random Forests and Boosted Trees are used across industry with great success; adding a level of insight simultaneously boosts model effectiveness and consumer trust.

TL;DR: We present a method for deriving instance-level explanations for tree ensemble models and examine its applications.

Keywords: Random Forests, Boosted Trees, Intelligible Models, Explainability, Interpretability, Tree Ensembles