Shedding Light on the Asymmetric Learning Capability of AdaBoost
This addresses a problem for machine learning practitioners by clarifying AdaBoost's capabilities, though it appears incremental as it reinterprets an existing algorithm.
The paper tackles the misconception that AdaBoost cannot handle asymmetric learning by proposing a new analysis that shows it can be used directly as an asymmetric algorithm while preserving theoretical properties, presenting a novel class-conditional description to model this behavior.
In this paper, we propose a different insight to analyze AdaBoost. This analysis reveals that, beyond some preconceptions, AdaBoost can be directly used as an asymmetric learning algorithm, preserving all its theoretical properties. A novel class-conditional description of AdaBoost, which models the actual asymmetric behavior of the algorithm, is presented.