Multivariate Probability Assessment.
MASSACHUSETTS INST OF TECH CAMBRIDGE OPERATIONS RESEARCH CENTER
Pagination or Media Count:
In order that formal Bayesian decision analysis may be applied to a specific decision problem, probabilities must be assigned to the associated uncertain quantities. Very often, the only available, relevant information concerning these uncertain quantities probabilities is an experts opinion, which may be represented mathematically by what is known as a judgmental probability. The assessment of judgmental probabilities involves numerous psychological and analytical problems. The first part of this report provides a concise, well referenced summary of previous work on the assessment of judgmental probabilities involving one uncertain quantity. The rest of the paper directly concerns the development of techniques for multivariate, judgmental probability assessment, i.e., assessing a joint probability distribution of two or more uncertain quantities. A major portion of this research investigates the use of the mutual probabilistic independence property in multivariate probability assessment. Also, a technique for the assessment of two dependent, uncertain quantities is developed. The method derives a representation of the desired joint probability density function, in the form of a set of two dimensional slices of its surface, from assessed marginal and cumulative distributions. Author
- Statistics and Probability
- Operations Research