A Boosting Algorithm for Label Covering in Multilabel Problems
Yonatan Amit, Ofer Dekel, Yoram Singer · 2007
We describe, analyze and experiment with a boosting algorithm for multilabel catego-rization problems. Our algorithm includes as special cases previously studied boosting al-gorithms such as Adaboost.MH. We cast the multilabel problem as multiple binary deci-sion problems, based on a user-defined cov-ering of the set of labels. We prove a lower bound on the progress made by our algorithm on each boosting iteration and demonstrate the merits of our algorithm in experiments with text categorization problems. 1