Date of this Version
JMLR: Workshop and Conference Proceedings
We consider the problem of forecasting a sequence of outcomes from an unknown source. The quality of the forecaster is measured by a family of checking rules. We prove upper bounds on the value of the associated game, thus certifying the existence of a calibrated strategy for the forecaster. We show that complexity of the family of checking rules can be captured by the notion of a sequential cover introduced in . Various natural assumptions on the class of checking rules are considered, including finiteness of Vapnik-Chervonenkis and Littlestone's dimensions.
Foster, D. P., Rakhlin, A., Sridharan, K., & Tewari, A. (2011). Complexity-Based Approach to Calibration With Checking Rules. JMLR: Workshop and Conference Proceedings, 19 293-314. Retrieved from https://repository.upenn.edu/statistics_papers/141
Date Posted: 27 November 2017
This document has been peer reviewed.