Online Boosting Algorithms for Multi-label Ranking
This work addresses multi-label learning for applications requiring label rankings, but it is incremental as it adapts boosting to an online setting.
The paper tackles the problem of multi-label ranking by designing online boosting algorithms with provable loss bounds, achieving performance at least as good as existing batch methods on real datasets.
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.