C.-S. Ferng & H.-T. Lin; JMLR
W&CP 20:281–295, 2011.
Multi-label Classiﬁcation with Error-correcting Codes
We formulate a framework for applying error-correcting codes (ECC) on multi-label
classiﬁcation problems. The framework treats some base learners as noisy channels and
uses ECC to correct the prediction errors made by the learners. An immediate use of
the framework is a novel ECC-based explanation of the popular random k
(RAKEL) algorithm using a simple repetition ECC. Using the framework, we empirically
compare a broad spectrum of ECC designs for multi-label classiﬁcation. The results
not only demonstrate that RAKEL can be improved by applying some stronger ECC,
but also show that the traditional Binary Relevance approach can be enhanced by
learning more parity-checking labels. In addition, our study on diﬀerent ECC helps
understand the trade-oﬀ between the strength of ECC and the hardness of the base learning
Page last modified on Sun Nov 6 15:43:59 2011.