On the interpretability of conditional probability estimates in the agnostic setting (Q1688991)

The authors study a binary classification problem based on conditional probability estimates. In this framework, they assume agnostic setting meaning that there is no knowledge about the underlying distribution. Under the agnostic setting, conditional probability estimates do not necessarily reflect the true conditional probabilities. Instead, they have a certain calibration property: among all data points that the classifier has predicted \(P(Y =1 |X)=p,\) \(p\) portion of them actually have label \(Y = 1\). For cost-sensitive decision problems, this calibration property provides adequate support for using Bayes Decision Rule. \newline The authors propose a novel measure for the calibration property together with its empirical counterpart, and prove a uniform convergence result between them. This new measure can be used to estimate the expected cost of decision rules when applied to an unlabeled dataset.