Modeling Skewed Class Distributions by Reshaping the Concept Space

We introduce an approach to learning from imbalanced class distributions that does not change the underlying data distribution. The ICC algorthm decomposes majority classes into smaller subclasses that create a more balanced class distribution. In this paper, we explain how ICC can not only address the class imbalance problem but may also increase the expressive power of the hypothesis space. We validate ICC and analyze alternative decomposition methods on well-known machine learning datasets as well as new problems in pervasive computing. Our results indicate that ICC performs as well or better than existing approaches to handling class imbalance.