Skoči na glavni sadržaj

Stručni rad

Data clustering

Kristian Sabo ; Odjel za matematiku,Sveučilište J. J. Strossmayera u Osijeku, Osijek, Hrvatska
Rudolf Scitovski ; Odjel za matematiku,Sveučilište J. J. Strossmayera u Osijeku, Osijek, Hrvatska
Ivan Vazler ; Odjel za matematiku,Sveučilište J. J. Strossmayera u Osijeku, Osijek, Hrvatska


Puni tekst: hrvatski pdf 1.671 Kb

str. 149-178

preuzimanja: 1.732

citiraj


Sažetak

In this paper we consider a clustering problem for a data-points set
$\mathcal{A}$ into disjoint nonempty subsets - clusters, whereby
it is assumed that elements of the set $\mathcal{A}$ are determined by
one or two characteristics. Least square criteria and least absolute deviation criteria are used for
solving the problem. A number of examples illustrating differences
between these criteria are given. Corresponding software support is
developed for the purpose of facilitating scientific or professional
work by using this methodology and approach.

Ključne riječi

clusters, arithmetic mean, median, optimization

Hrčak ID:

66979

URI

https://hrcak.srce.hr/66979

Podaci na drugim jezicima: hrvatski

Posjeta: 2.419 *