Everything you always wanted to know about normalization (but were afraid to ask)
Abstract
The solution to the problem of normalization of variables with different units of measurement is of primary interest in data analysis. Most of the reports in the literature present a wide variety of normalization methods, but do not explain how to choose the ‘right’ method. Researchers cannot avoid the question simply by choosing the normalization on the basis of how they want to present the results, as each method has its pros and cons. In this paper, a comparison among the main normalization methods is presented and an alternative method, denoted as ‘Re-scaling with a reference’ is proposed. Some issues on the ‘effect’ of normalization and suggestions for a correct choice of the normalization method are also reported.
Downloads
Published
Issue
Section
License
Copyright (c) 2021 Matteo Mazziotta, Adriano Pareto
This work is licensed under a Creative Commons Attribution 4.0 International License.