A Family of Additive Online Algorithms for Category Ranking
Koby Crammer, Yoram Singer;
We describe a new family of topic-ranking algorithms for multi-labeled
documents. The motivation for the algorithms stem from recent advances
in online learning algorithms. The algorithms are simple to implement
and are also time and memory efficient. We provide a unified analysis
of the family of algorithms in the mistake bound model. We then discuss
experiments with the proposed family of topic-ranking algorithms on the
Reuters-21578 corpus and the new corpus released by Reuters in 2000.
On both corpora, the algorithms we present achieve
results and outperforms topic-ranking adaptations of Rocchio's algorithm
and of the Perceptron algorithm.