Accession Number : AD0681029

Title :   STATISTICAL TECHNIQUE IN CLUSTERING AND PATTERN RECOGNITION.

Descriptive Note : Technical rept.,

Corporate Author : STANFORD UNIV CALIF DEPT OF STATISTICS

Personal Author(s) : Switzer,Paul

Report Date : 18 OCT 1968

Pagination or Media Count : 29

Abstract : Several techniques in clustering and pattern recognition are examined and evaluated. In each the objective is to partition the elements of a specified population into subsets so that the partition meets conditions imposed by the problems. The basic metric employed is the distance between elements. The question of sampling techniques in clustering is investigated because total enumeration of partitions very quickly becomes inordinately high. Several specific examples are worked out. (Author)

Descriptors :   (*PATTERN RECOGNITION, *COMBINATORIAL ANALYSIS), SAMPLING, INFORMATION THEORY, DISTRIBUTION THEORY, PROBABILITY

Subject Categories : Statistics and Probability
      Bionics

Distribution Statement : APPROVED FOR PUBLIC RELEASE