Pattern Normalization – a New Tool for Dynamic Comparisons

Iwona Müller-Frączek

Pattern Normalization – a New Tool for Dynamic Comparisons

Číslo: 2/2019
Periodikum: Statistika

Klíčová slova: Normalization, standardization, transformation of variables

Pro získání musíte mít účet v Citace PRO.

Přečíst po přihlášení

Anotace: The article presents a new method of normalization – normalization with respect to pattern (or pattern

normalization in short). It has properties expected for this type of transformation: preserves skewness, kurtosis
and the Pearson correlation coefficients. Although pattern normalization uses only observations from the current
unit of time, it can be used in dynamic research. An additional advantage of new normalization is the ability
to reflect different analysis environments. The effects of pattern normalization are illustrated by an empirical example. Indicators monitoring the implementation of the Europe 2020 Strategy are used. Normalizations are carried out for two reference groups: the entire EU and countries that joined the EU in 2004. The results for two years are compared. The example of Poland shows that the “dynamic image” of the country is affected by the use of pattern normalization itself as well as by the choice of the environment. In this context pattern normalization is similar to dynamic standardization, and different than dynamic scaling.