Accession Number : AD0693144

Title :   A NONPARAMETRIC APPROACH TO PATTERN RECOGNITION. PART II. THE NON-DISJOINT CASE.

Descriptive Note : Technical rept.,

Corporate Author : INFORMATION RESEARCH ASSOCIATES INC WALTHAM MASS

Personal Author(s) : Owen,Joel ; Henrichon,Ernest G. , Jr

Report Date : 04 OCT 1968

Pagination or Media Count : 19

Abstract : In Part I of this paper, (AD-664 218), a nonparametric discrimination technique was proposed. It was shown there that when perfect discrimination was possible, this technique achieved perfection and in certain cases achieved it with a finite learning phase. In this report, the technique is modified to include the case when perfect discrimination is not possible. It is shown that this procedure yields results which converge in probability to the optimal decision boundaries determined by the likelihood ratio method. (Author)

Descriptors :   (*PATTERN RECOGNITION, DECISION THEORY), INFORMATION THEORY, LEARNING MACHINES, ARTIFICIAL INTELLIGENCE, STATISTICAL ANALYSIS, PROBABILITY, ALGORITHMS

Subject Categories : Cybernetics
      Bionics

Distribution Statement : APPROVED FOR PUBLIC RELEASE