Loading…

PRELIMINARY NOTE ON THE CLASSIFICATION OF VECTORS WITH SOME UNKNOWN DESCRIPTORS

The problem of pattern classification based upon incomplete vectors was introduced in an earlier paper, where we discussed four possible methods of solution. Of these, only two appeared to be promising as a solution to the general problem. In one of them (a), we envisaged that a multi-dimensional Mo...

Full description

Saved in:
Bibliographic Details
Published in:Kybernetes 1975-02, Vol.4 (2), p.105-109
Main Authors: BATCHELOR, B.G, HAND, D.J
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The problem of pattern classification based upon incomplete vectors was introduced in an earlier paper, where we discussed four possible methods of solution. Of these, only two appeared to be promising as a solution to the general problem. In one of them (a), we envisaged that a multi-dimensional Monte-Carlo integration would be needed, while the other procedure (b) required the use of a large store. In this paper, we show how the long integration may be avoided in procedure (a). This is feasible if we use one of a certain class of p.d.f. estimators. We also show that these p.d.f. estimators have certain desirable properties in Method (b).
ISSN:0368-492X
DOI:10.1108/eb005385