Work

Adaptive calibration for binary classification.

Year: 2021
Type: preprint
Abstract: This note proposes a way of making probability forecasting rules less sensitive to changes in data distribution, concentrating on the simple case of binary classification. This is important in applications of machine learning, where the quality of a trained predictor may drop significantly in the process of its exploitation. Our techniques are based on recent work on conformal test martingales and older work on prediction with expert advice, namely tracking the best expert.
Language: English
Open Access status: green
OpenAlex ID: W3181339911
Cites: 10
Cited by:
Related to: 10