The Concept of a Best Approximation as an Optimality Criterion in Statistical Decision Theory.
Abstract
The concept of a best approximation from a convex set in a normed linear space is interpreted as a natural choice for a concept of optimality in statistical decision. The fundamental distinction for statistics is that there is a specific orientation to the approximation induced by the notion of admissibility in statistical decision theory.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1975
- Accession Number
- ADA020213
Entities
People
- Bernard Harris
- Gerhard Heindl
Organizations
- University of Wisconsin–Madison