GENERATION OF POLYNOMIAL DISCRIMINANT FUNCTIONS FOR PATTERN RECOGNITION.

Abstract

The purpose of this research is to derive a simple method of determining weights for crossproduct and power terms in the variable inputs to an adaptive threshold element used for statistical pattern classification. The broad objective is to make it possible to realize general nonlinear decision surfaces, in contrast with the linear (hyperplanar) decision surfaces that can be realized by an adaptive threshold element using only first-order terms as inputs. The derivation is based on nonparametric estimation of a probability density function for each category to be classified so that the Bayes decision rule can be used for classification. The derivation has been carried out in such a way that the decision surfaces which separate categories have good extrapolating ability even when the number of training patterns is quite small. The primary contribution of this research is the development and analysis of the polynomial discriminant method (PDM) of pattern recognition by which nonlinear decision surfaces can be established in a way which is both theoretically well-founded and decidedly practical to implement. The basic PDM algorithms are designed for use with analog measurement variables; a simplified PDM algorithm has also been derived for binary variables. Implementation of the PDM, both in the form of computer programs and in the form of polynomial threshold devices, is discussed.

Document Details

Document Type
Technical Report
Publication Date
May 01, 1966
Accession Number
AD0487537

Entities

People

  • Donald F. Specht

Organizations

  • Stanford University

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Algorithms
  • Classification
  • Computer Programs
  • Computers
  • Contrast
  • Demographic Cohorts
  • Mathematics
  • Measurement
  • Pattern Recognition
  • Polynomials
  • Probability
  • Probability Density Functions
  • Recognition
  • Training

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms