DeepAI AI Chat
Log In Sign Up

Purifying Interaction Effects with the Functional ANOVA: An Efficient Algorithm for Recovering Identifiable Additive Models

by   Benjamin Lengerich, et al.

Recent methods for training generalized additive models (GAMs) with pairwise interactions achieve state-of-the-art accuracy on a variety of datasets. Adding interactions to GAMs, however, introduces an identifiability problem: effects can be freely moved between main effects and interaction effects without changing the model predictions. In some cases, this can lead to contradictory interpretations of the same underlying function. This is a critical problem because a central motivation of GAMs is model interpretability. In this paper, we use the Functional ANOVA decomposition to uniquely define interaction effects and thus produce identifiable additive models with purified interactions. To compute this decomposition, we present a fast, exact, mass-moving algorithm that transforms any piecewise-constant function (such as a tree-based model) into a purified, canonical representation. We apply this algorithm to several datasets and show large disparity, including contradictions, between the apparent and the purified effects.


GAMI-Net: An Explainable Neural Network based on Generalized Additive Models with Structured Interactions

The lack of interpretability is an inevitable problem when using neural ...

Constrained functional additive models for estimating interactions between a treatment and functional covariates

A novel functional additive model is proposed which is uniquely modified...

The SKIM-FA Kernel: High-Dimensional Variable Selection and Nonlinear Interaction Discovery in Linear Time

Many scientific problems require identifying a small set of covariates t...

Using Model-Based Trees with Boosting to Fit Low-Order Functional ANOVA Models

Low-order functional ANOVA (fANOVA) models have been rediscovered in the...

Moving from two- to multi-way interactions among binary risk factors on the additive scale

Many studies have focused on investigating deviations from additive inte...

Unifying local and global model explanations by functional decomposition of low dimensional structures

We consider a global explanation of a regression or classification funct...

Understanding and Exploiting Object Interaction Landscapes

Interactions play a key role in understanding objects and scenes, for bo...