A Formalism to Generate Probability Distributions for Performance-Assessment Modeling

by Paul G. Kaplan, Sandia Natl Lab, Albuquerque, United States,

Abstract: A formalism is presented for generating probability distributions of parameters used in performance-assessment modeling. The formalism is used when data are either sparse or nonexistent. The appropriate distribution is a function of the known or estimated constraints and is chosen to maximize a quantity known as Shannon's informational entropy. The formalism is applied to a parameter used in performance-assessment modeling. The functional form of the model that defines the parameter, data from the actual field site, and natural analog data are analyzed to estimate the constraints. A beta probability distribution of the example parameter is generated after finding four constraints.

Subject Headings: Probability | Probability distribution | Parameters (statistics) | Entropy methods | Data processing | Data analysis | Site investigation

Services: Buy this book/Buy this article


Return to search