CLASSIFICATION BY STATISTICAL METHODS (PATTERN RECOGNITION)
Abstract
Mathematical methods of pattern recognition which make use of the theory of statistical inference are treated. Decision theoretic methods of classification, and learning through parameter estimation form the foundation for this discussion. Although the procedures are formulated in terms of a general functional representation of the distributions, specific solutions are outlined for multivariate Normal distributions. This is followed with a brief commentary on the potential contribution from the field of nonparametric statistics. The analysis concludes with a list of potential applications. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1961
- Accession Number
- AD0278689
Entities
People
- P.w. Cooper
Organizations
- Melpar