Decision Theory for Averaged Risk,
Abstract
Let Phi denote a decision rule with risk R (theta, Phi) depending on a parameter theta. A new risk function called the 'averaged risk' is obtained by averaging R(theta, Phi) with respect to a probability measure G sub phi on the parameter space. The class of admissible, Bayes and minimax rules with respect to the new risk function is considered and compared with the corresponding class of rules with respect to R (theta, Phi). (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 15, 1971
- Accession Number
- AD0736170
Entities
People
- David Albert Ott
Organizations
- Clemson University