Wybór grup metod normalizacji wartości zmiennych w skalowaniu wielowymiarowym
The Choice of Groups of Variable Normalization Methods in Multidimensional Scaling
Author(s): Marek WalesiakSubject(s): Economy, Tourism, ICT Information and Communications Technologies
Published by: Główny Urząd Statystyczny
Keywords: normalization of variables; multidimensional scaling; distance measures; R program; clusterSim package
Summary/Abstract: In multidimensional scaling carried out on the basis of metric data matrix (interval, ratio) one of the stages is the choice of the variable normalization method. The R package clusterSim with data.Normalization function has been developed for that purpose. It provides 18 data normalization methods.In this paper the proposal of procedure which allows to isolate groups of normalization methods that lead to similar multidimensional scaling results were presented. The proposal can reduce the problem of choosing the normalization method in multidimensional scaling. The results are illustrated via empirical example.
Journal: Przegląd Statystyczny. Statistical Review
- Issue Year: 63/2016
- Issue No: 1
- Page Range: 7-18
- Page Count: 12
- Language: Polish