Title: A nonlinear multivariate method for constructing profiles of country performance
Authors: Sara Casacci; Adriano Pareto
Addresses: Italian National Institute of Statistics, Rome, Italy ' Italian National Institute of Statistics, Rome, Italy
Abstract: Composite indices for measuring multi-dimensional phenomena such as sustainable development, quality of life, and competitiveness, are based on individual indicators that have no common unit of measurement. As a rule, individual indicators are normalised by univariate methods, independently of each other. However, normalisation has a strong impact on results since it creates a 'correspondence system' among different variables. Thus the normalisation method should take into account the relationships existing in the data. In this paper, a nonlinear multivariate method is proposed for normalising a set of individual indicators in order to aggregate them into a composite index and construct profiles of country performance. An application to the measurement and analysis of socio-economic development is also shown.
Keywords: composite index; indicators; normalisation; nonlinear relationship.
International Journal of Multicriteria Decision Making, 2021 Vol.8 No.4, pp.295 - 312
Received: 31 Mar 2021
Accepted: 04 Jun 2021
Published online: 07 Feb 2022 *