THREE THEORETICAL SCHEMES OF LEARNING RECOGNITION SYSTEMS,
Abstract
A perfect perceptron-type learning pattern recognizer function is described. A real recognizer can be described by a function f(x, u). The learning problem includes the selection of the u-parameter in such a way that the approximation is as accurate as possible; the values of s(x) are known only at given points.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 04, 1969
- Accession Number
- AD0699232
Entities
People
- V. A. Yakubovich
Organizations
- National Air and Space Intelligence Center