Efficient distributional regression trees learning algorithms for calibrated non-parametric probabilistic forecasts

TL;DR

This paper presents efficient algorithms for training probabilistic regression trees that optimize proper scoring rules like WIS and CRPS, enhancing uncertainty estimation in non-parametric regression.

Contribution

It introduces novel, computationally efficient algorithms for distributional regression trees based on proper scoring rules, improving probabilistic forecasts.

Findings

Algorithms are computationally efficient using known data structures.

Methods achieve competitive performance with existing approaches.

Trees provide interpretable models with group-conditional coverage guarantees.

Abstract

The perspective of developing trustworthy AI for critical applications in science and engineering requires machine learning techniques that are capable of estimating their own uncertainty. In the context of regression, instead of estimating a conditional mean, this can be achieved by producing a predictive interval for the output, or to even learn a model of the conditional probability $p(y|x)$ of an output $y$ given input features $x$. While this can be done under parametric assumptions with, e.g. generalized linear model, these are typically too strong, and non-parametric models offer flexible alternatives. In particular, for scalar outputs, learning directly a model of the conditional cumulative distribution function of $y$ given $x$ can lead to more precise probabilistic estimates, and the use of proper scoring rules such as the weighted interval score (WIS) and the continuous ranked probability score (CRPS) lead to better coverage and calibration properties. This paper introduces novel algorithms for learning probabilistic regression trees for the WIS or CRPS loss functions. These algorithms are made computationally efficient thanks to an appropriate use of known data structures - namely min-max heaps, weight-balanced binary trees and Fenwick trees. Through numerical experiments, we demonstrate that the performance of our methods is competitive with alternative approaches. Additionally, our methods benefit from the inherent interpretability and explainability of trees. As a by-product, we show how our trees can be used in the context of conformal prediction and explain why they are particularly well-suited for achieving group-conditional coverage guarantees.