PLATTERN CLASSIFICATION AND SWITCHING THEORY.

Abstract

This report is concerned with certain aspects of pattern classification and their relation to switching theory. Attention is focused on pattern classification problems involving two classes. A pattern classification problem consists of three fundamental aspects, namely characterization or the choice of the independent variables of the problem, the so-called primary attributes, the abstraction or the determination of a decision surface based on a certain number of given samples of either class, and generalization or the ability of the decision surface to classify correctly a sample whose class is unknown. (Author)

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1965
Accession Number
AD0473537

Entities

People

  • R. L. Kashyap

Organizations

  • Harvard University

Tags

DTIC Thesaurus Topics

  • Classification

Readers

  • Neural Network Machine Learning.
  • Theoretical Analysis.