A Kullback-Leibler View of Linear and Log-Linear Pools

Authors:
Ali E. Abbas
Affiliations:
Department of Industrial and Enterprise Systems Engineering, College of Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801
Venue:
Decision Analysis
Year:
2009

Citing 4
Cited 12

Elements of information theory

Elements of information theory
Elicitation of Probabilities Using Competitive Scoring Rules

Decision Analysis
Some Comparisons among Quadratic, Spherical, and Logarithmic Scoring Rules

Decision Analysis
The Parimutuel Kelly Probability Scoring Rule

Decision Analysis

From the Editor...

Decision Analysis
From the Editors...

Decision Analysis
Honoring Michael H. Rothkopf's Legacy of Rigor and Relevance in Auction Theory: From the Editors

Decision Analysis
Managing Project Failure Risk Through Contingent Contracts in Procurement Auctions

Decision Analysis
From the Editors...

Decision Analysis
From the Editors...

Decision Analysis
Combining marginal probability distributions via minimization of weighted sum of Kullback--Leibler divergences

International Journal of Approximate Reasoning
From the Editors---Deterrence, Multiattribute Utility, and Probability and Bayes' Updating

Decision Analysis
From the Editor---Multiattribute and Intertemporal Preferences, Probability, and Stochastic Processes: Models and Assessment

Decision Analysis
From the Editors---Probability Scoring Rules, Ambiguity, Multiattribute Terrorist Utility, and Sensitivity Analysis

Decision Analysis
Game-theoretic probability combination with applications to resolving conflicts between statistical methods

International Journal of Approximate Reasoning
From the Editors---Games and Decisions in Reliability and Risk

Decision Analysis

Quantified Score

Hi-index	0.00

Visualization

Abstract

Linear and log-linear pools are widely used methods for aggregating expert belief. This paper frames the expert aggregation problem as a decision problem with scoring rules. We propose a scoring function that uses the Kullback-Leibler (KL) divergence measure between the aggregate distribution and each of the expert distributions. The asymmetric nature of the KL measure allows for a convenient scoring system for which the linear and log-linear pools provide the optimal assignment. We also propose a “goodness-of-fit” measure that determines how well each opinion pool characterizes its expert distributions, and also determines the performance of each pool under this scoring function. We work through several examples to illustrate the approach.