Multiclass boosting with asymmetric binary weaklearners.
We introduce a multiclass generalization of AdaBoost with binary weaklearners.
We use a vectorial codification to represent class labels and a multiclass
exponential loss function to evaluate classifier responses. This representation
produces a set of margin values that provide a range of punishments for failures
and rewards for successes. Moreover, the stagewise optimization of this model
introduces an asymmetric boosting procedure whose costs depend on the number
of classes separated by each weaklearner. In this way the boosting procedure
takes into account class imbalances when building the ensemble. The experiments
performed compare this new approach favorably to AdaBoost.MH, GentleBoost and
the SAMME algorithms.
Related Publications:
PR'2014

Multiclass costsensitive boosting: BAdaCost.
We introduce a costsensitive multiclass Boosting algorithm (BAdaCost) based on a generalization
of the Boosting margin, termed multiclass costsensitive margin. It can be used, for example, to
address the class imbalance by the introduction of a cost matrix that weighs more hevily the costs
of confused classes.
Related Publications:
IBPRIA'2015
