Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Is there a neat way to use something like the Brier score to score an infinite set of forecasts/outcomes?

share|improve this question
add comment

1 Answer 1

up vote 5 down vote accepted

A general form of the Brier score, for an essentially arbitrary outcome space $\cal X$, is as follows.

Let $q(\cdot)$ be your quoted density for a random quantity $X$, with respect to a dominating measure $\mu$ over $\cal X$. Then your score, when outcome $X=x$ is realised, is $$S(x, q(\cdot)) = \int q(t)^2 d\mu(t) - 2q(x).$$

The Brier score is just one of an infinity of proper scoring rules. For some alternatives, see e.g.:

Gneiting, T. and Raftery, A. E. (2007). Strictly proper scoring rules, prediction, and estimation. Journal of the American Statistical Association, 102, 359-378.
Dawid, A. P. (2007). The geometry of proper scoring rules. Annals of the Institute of Statistical Mathematics 59, 77-93.
doi:10.1007/s10463-006-0099-8

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.