Online Boosting Algorithms for Multi-label Ranking

TL;DR

This paper introduces online boosting algorithms tailored for multi-label ranking tasks, providing provable guarantees and demonstrating competitive performance against existing batch methods on real datasets.

Contribution

The paper presents the first online boosting algorithms for multi-label ranking with theoretical loss bounds and practical adaptive variants.

Findings

Algorithms achieve optimal learner requirements for accuracy

Adaptive algorithm does not need knowledge of weak learner edges

Experimental results show competitive performance with batch methods

Abstract

We consider the multi-label ranking approach to multi-label learning. Boosting is a natural method for multi-label ranking as it aggregates weak predictions through majority votes, which can be directly used as scores to produce a ranking of the labels. We design online boosting algorithms with provable loss bounds for multi-label ranking. We show that our first algorithm is optimal in terms of the number of learners required to attain a desired accuracy, but it requires knowledge of the edge of the weak learners. We also design an adaptive algorithm that does not require this knowledge and is hence more practical. Experimental results on real data sets demonstrate that our algorithms are at least as good as existing batch boosting algorithms.