PRIMA

a new pattern recognition method

I. Juricskay, Gábor E. Veress

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

PRIMA, a new supervised classification method is based on the concept of class distance (Euclidean distance). For each class, a separate class distance is defined on the basis of the centre of gravity and inhomogeneity for the class; this class distance is then used to produce the classification. The PRIMA classifier based on class distance can be applied in different, complex cases. The conditions of applicability are less strict than those of other methods. The algorithm is simple; efficiency and stability are good. The simplicity of the method even for complex cases, e.g., with very many variables or multicategory classification, or with noisy or incomplete data processing, is noteworthy when compared with other effective pattern recognition methods.

Original languageEnglish
Pages (from-to)61-76
Number of pages16
JournalAnalytica Chimica Acta
Volume171
Issue numberC
DOIs
Publication statusPublished - May 1 1985

Fingerprint

pattern recognition
Pattern recognition
Gravitation
Classifiers
image classification
inhomogeneity
gravity
method

ASJC Scopus subject areas

  • Biochemistry
  • Analytical Chemistry
  • Spectroscopy
  • Environmental Chemistry

Cite this

PRIMA : a new pattern recognition method. / Juricskay, I.; Veress, Gábor E.

In: Analytica Chimica Acta, Vol. 171, No. C, 01.05.1985, p. 61-76.

Research output: Contribution to journalArticle

Juricskay, I. ; Veress, Gábor E. / PRIMA : a new pattern recognition method. In: Analytica Chimica Acta. 1985 ; Vol. 171, No. C. pp. 61-76.
@article{d12ab5d55bac47f8b9b6ea6b7547f067,
title = "PRIMA: a new pattern recognition method",
abstract = "PRIMA, a new supervised classification method is based on the concept of class distance (Euclidean distance). For each class, a separate class distance is defined on the basis of the centre of gravity and inhomogeneity for the class; this class distance is then used to produce the classification. The PRIMA classifier based on class distance can be applied in different, complex cases. The conditions of applicability are less strict than those of other methods. The algorithm is simple; efficiency and stability are good. The simplicity of the method even for complex cases, e.g., with very many variables or multicategory classification, or with noisy or incomplete data processing, is noteworthy when compared with other effective pattern recognition methods.",
author = "I. Juricskay and Veress, {G{\'a}bor E.}",
year = "1985",
month = "5",
day = "1",
doi = "10.1016/S0003-2670(00)84184-7",
language = "English",
volume = "171",
pages = "61--76",
journal = "Analytica Chimica Acta",
issn = "0003-2670",
publisher = "Elsevier",
number = "C",

}

TY - JOUR

T1 - PRIMA

T2 - a new pattern recognition method

AU - Juricskay, I.

AU - Veress, Gábor E.

PY - 1985/5/1

Y1 - 1985/5/1

N2 - PRIMA, a new supervised classification method is based on the concept of class distance (Euclidean distance). For each class, a separate class distance is defined on the basis of the centre of gravity and inhomogeneity for the class; this class distance is then used to produce the classification. The PRIMA classifier based on class distance can be applied in different, complex cases. The conditions of applicability are less strict than those of other methods. The algorithm is simple; efficiency and stability are good. The simplicity of the method even for complex cases, e.g., with very many variables or multicategory classification, or with noisy or incomplete data processing, is noteworthy when compared with other effective pattern recognition methods.

AB - PRIMA, a new supervised classification method is based on the concept of class distance (Euclidean distance). For each class, a separate class distance is defined on the basis of the centre of gravity and inhomogeneity for the class; this class distance is then used to produce the classification. The PRIMA classifier based on class distance can be applied in different, complex cases. The conditions of applicability are less strict than those of other methods. The algorithm is simple; efficiency and stability are good. The simplicity of the method even for complex cases, e.g., with very many variables or multicategory classification, or with noisy or incomplete data processing, is noteworthy when compared with other effective pattern recognition methods.

UR - http://www.scopus.com/inward/record.url?scp=0344522439&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0344522439&partnerID=8YFLogxK

U2 - 10.1016/S0003-2670(00)84184-7

DO - 10.1016/S0003-2670(00)84184-7

M3 - Article

VL - 171

SP - 61

EP - 76

JO - Analytica Chimica Acta

JF - Analytica Chimica Acta

SN - 0003-2670

IS - C

ER -