Some General Principles for the Dual Problem to Statistical Classification.
Abstract
We consider the design of decision problems which maximize the classification error for a given set of discriminants. A minimax principle is proved, which has applications in discriminant analysis and feature extraction. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 26, 1980
- Accession Number
- ADA094726
Entities
People
- Lee K. Jones
Organizations
- Massachusetts Institute of Technology