Kesra Nermend, Ph., D., Institute of IT in Management, University of Szczecin, Mickiewicza 64, Szczecin, 710101, Poland
In the analysis of the economic objects, the problem of outliers appears. The presence of outstanding observations causes that classes that are constructed applying taxonomic measures are not homogeneous. Thus it is necessary to recognize such objects and minimize their influence.
In order to solve that problem we use quartiles for the synthetic taxonomic measure construction. The aim of the paper is presentation of the idea of the proposed method and to discuss the advantages of the quartiles introducing to the particular application.