Eliciting continuous probability distributions

The process of eliciting a representative probability distribution from a judge is fundamental to decision making under uncertainty. This work summarizes several aspects of eliciting continuous probability distributions. Through experimentation involving 103 judges the method of assessing single variable distributions by fixed variable or fixed probability method is compared. Slight, but consistent, superiority is shown by the fixed variable method. The use of iso-probability contours, a novel methodology for assessing joint probability distributions, is tested through experimentation. The ability of this method to assess a judges belief about the correlation of two variables is characterized. Further recommendations for the use of iso-probability contours are provided. The variable weighting function is introduced. This novel tool is designed to represent a subjects true belief about a continuous probability distribution as explained by cumulative prospect theory. Operating similarly to the probability weighting function, the variable weighting function allows an assessor to gain a judges belief about a specific variable value without requiring multiple assessments. Functional form of the variable weighting function is explored and relevant parameter values are provided.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Systems & Entrepreneurial Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2011

Author dc:creator

• Haggerty, Ryan P.

Contributors dc:contributor

• Abbas, Ali E.

• Decision Analysis
• Probability Assessment
• Cumulative Prospect Theory

Statement dc:rights

• Copyright 2011 Ryan Haggerty

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/24099

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/24099