Knowledge Integration for Conditional Probability Assessments
This work provides incremental theoretical insights for researchers in probabilistic uncertainty management.
The paper addresses the problem of integrating two discrete conditional probability distributions by analyzing their consistency and deriving coherence conditions and explicit formulas for extending to marginal distributions in special cases.
In the probabilistic approach to uncertainty management the input knowledge is usually represented by means of some probability distributions. In this paper we assume that the input knowledge is given by two discrete conditional probability distributions, represented by two stochastic matrices P and Q. The consistency of the knowledge base is analyzed. Coherence conditions and explicit formulas for the extension to marginal distributions are obtained in some special cases.