Proper local scoring rules on discrete sample spaces

DOI10.1214/12-AOS972zbMATH Open1246.62010arXiv1104.2224WikidataQ57394014 ScholiaQ57394014MaRDI QIDQ450050FDOQ450050

Authors: A. Philip Dawid, Steffen Lauritzen, Matthew Parry 

Publication date: 3 September 2012

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: A scoring rule is a loss function measuring the quality of a quoted probability distribution $Q$ for a random variable $X$, in the light of the realized outcome $x$ of $X$; it is proper if the expected score, under any distribution $P$ for $X$, is minimized by quoting $Q=P$. Using the fact that any differentiable proper scoring rule on a finite sample space $mathcalX$ is the gradient of a concave homogeneous function, we consider when such a rule can be local in the sense of depending only on the probabilities quoted for points in a nominated neighborhood of $x$. Under mild conditions, we characterize such a proper local scoring rule in terms of a collection of homogeneous functions on the cliques of an undirected graph on the space $mathcalX$. A useful property of such rules is that the quoted distribution $Q$ need only be known up to a scale factor. Examples of the use of such scoring rules include Besag's pseudo-likelihood and Hyv"{a}rinen's method of ratio matching.

Full work available at URL: https://arxiv.org/abs/1104.2224

Recommendations

• Proper local scoring rules
• Theory and applications of proper scoring rules
• Local proper scoring rules of order two
• The uniqueness of local proper scoring rules: the logarithmic family
• Extensive scoring rules

zbMATH Keywords

entropyconcavityEuler's theoremhomogeneous functionSsupergradient

Mathematics Subject Classification ID

Applications of graph theory (05C90) Statistical decision theory (62C99)

Cites Work

• Predictive model assessment for count data
• Title not available (Why is that?)
• Convex Analysis
• Some extensions of score matching
• Expected information as ecpected utility
• Robust estimation in truncated discrete distributions with application to capture-recapture experiments
• MEASURES OF THE VALUE OF INFORMATION
• Proper local scoring rules
• Local proper scoring rules of order two
• Proper Scores for Probability Forecasters
• A Theorem about Random Fields
• The geometry of proper scoring rules

Cited In (20)

• A Novel Parameter Estimation Method for Boltzmann Machines
• Extensive scoring rules
• Estimation of spatial processes using local scoring rules: spatial special issue
• Weighted Scoring Rules and Convex Risk Measures
• Local proper scoring rules of order two
• Proper local scoring rules
• The Hyvärinen scoring rule in Gaussian linear time series models
• A note on Bayesian model selection for discrete data using proper scoring rules
• Bayesian model comparison with the Hyvärinen score: computation and consistency
• Prediction in several conventional contexts
• Affine invariant divergences associated with proper composite scoring rules and their applications
• Generalized Bayesian Inference for Discrete Intractable Likelihood
• Theory and applications of proper scoring rules
• Additive scoring rules for discrete sample spaces
• Objective Bayesian inference with proper scoring rules
• Title not available (Why is that?)
• Graph-based composite local Bregman divergences on discrete sample spaces
• The uniqueness of local proper scoring rules: the logarithmic family
• A note on the use of empirical AUC for evaluating probabilistic forecasts
• Bayesian experimental design without posterior calculations: an adversarial approach

This page was built for publication: Proper local scoring rules on discrete sample spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q450050)