Multi-label Chaining with Imprecise Probabilities

TL;DR

This paper extends multi-label chaining to handle imprecise probability estimates using credal sets, enabling cautious predictions and improved accuracy on uncertain instances, with efficient adaptations to the naive credal classifier.

Contribution

It introduces two strategies for multi-label chaining with imprecise probabilities and adapts them to the naive credal classifier, enhancing cautiousness and reliability.

Findings

Approaches produce relevant cautious predictions on hard instances.

Strategies are computationally efficient.

Imprecise models outperform precise ones on uncertain data.

Abstract

We present two different strategies to extend the classical multi-label chaining approach to handle imprecise probability estimates. These estimates use convex sets of distributions (or credal sets) in order to describe our uncertainty rather than a precise one. The main reasons one could have for using such estimations are (1) to make cautious predictions (or no decision at all) when a high uncertainty is detected in the chaining and (2) to make better precise predictions by avoiding biases caused in early decisions in the chaining. We adapt both strategies to the case of the naive credal classifier, showing that this adaptations are computationally efficient. Our experimental results on missing labels, which investigate how reliable these predictions are in both approaches, indicate that our approaches produce relevant cautiousness on those hard-to-predict instances where the precise models fail.