Construction of Prior Distributions

doi:10.1117/3.2540669.ch7

Book: Optimal Bayesian Classification

Author(s): Lori A. Dalton, Edward R. Dougherty

Published: 2020

https://doi.org/10.1117/3.2540669.ch7

Author Affiliations +

Abstract

Up to this point we have ignored the issue of prior construction, assuming that the characterization of uncertainty is known. For optimal Bayesian classification, the problem consists of transforming scientific knowledge into a probability distribution governing uncertainty in the feature-label distribution. Regarding prior construction in general, in 1968, E. T. Jaynes remarked, “Bayesian methods, for all their advantages, will not be entirely satisfactory until we face the problem of finding the prior probability squarely” (Jaynes, 1968). Twelve years later, he added, “There must exist a general formal theory of determination of priors by logical analysis of prior information—and that to develop it is today the top priority research problem of Bayesian theory” (Jaynes, 1980).