Exact Multidimensional Credibility.
Abstract
In a previous paper, it was shown that the linear exponential family of likelihoods, together with their natural conjugate priors, gave exact credibility in the one-dimensional case; that is, a Bayesian forecast of the mean observation which is linear in the data. This paper considers the multidimensional case, and the conditions under which exact credibility again holds.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1974
- Accession Number
- ADA002243
Entities
People
- William S. Jewell
Organizations
- University of California, Berkeley