Minimax Lower Bounds for Cost Sensitive Classification
Machine Learning
2018-05-22 v1 Machine Learning
Abstract
The cost-sensitive classification problem plays a crucial role in mission-critical machine learning applications, and differs with traditional classification by taking the misclassification costs into consideration. Although being studied extensively in the literature, the fundamental limits of this problem are still not well understood. We investigate the hardness of this problem by extending the standard minimax lower bound of balanced binary classification problem (due to \cite{massart2006risk}), and emphasize the impact of cost terms on the hardness.
Keywords
Cite
@article{arxiv.1805.07723,
title = {Minimax Lower Bounds for Cost Sensitive Classification},
author = {Parameswaran Kamalaruban and Robert C. Williamson},
journal= {arXiv preprint arXiv:1805.07723},
year = {2018}
}