A distance based classification method using an incomplete training set

  • Authors:

  • A Smolarz;M Usai;B Dubuisson

  • Affiliations:

  • University of Technology, ERA CNRS Heudiasyc, Department G.I., BP 233, 60206 Compiègne, France;University of Technology, ERA CNRS Heudiasyc, Department G.I., BP 233, 60206 Compiègne, France;University of Technology, ERA CNRS Heudiasyc, Department G.I., BP 233, 60206 Compiègne, France

  • Venue:
  • Pattern Recognition Letters
  • Year:
  • 1984

Quantified Score

Hi-index 0.10

Visualization

Abstract

This paper describes two distance based methods for classification when all the classes are not known. The first method is parametric, based on Gaussian assumption; the second one is nonparametric, based on the membership function concept.