Page 167 - Contributed Paper Session (CPS) - Volume 6
P. 167
CPS1866 Milica Maricic et al.
Multivariate approach to dimension
reduction based on the enhanced Scatter
Search – Composite I-distance indicator
(eSS-CIDI) approach: The case of the
Sustainable Society Index (SSI)
Milica Maricic, Veljko Jeremic, Milica Bulajic
University of Belgrade, Faculty of Organizational Sciences, Belgrade, Serbia
Abstract
Sustainability and sustainable development goals have become a major topic
of the world’s policy agenda. Nations worldwide are making efforts to create
sustainable societies for their citizens and future generations. However, the
issue of measuring the level of sustainability emerges. So far, composite
indicators have been used with success to provide decision-makers and wider
public the information regarding the achieved level of sustainability.
Nevertheless, frameworks of such composite indicators can be complex as
they incorporate indicators which measure different aspects of sustainability.
Therefore, herein we propose the application of the enhanced Scatter Search
– Composite I-distance indicator (eSS-CIDI) approach to reduce the number
of dimensions of a composite indicator. As a case study we chose the
acknowledged Sustainable Society Index (SSI). Our results show that the SSI
framework could be modified. The presented approach and obtained results
can be a foundation for further research on dimension reduction procedures
and composite indicators.
Keywords
Dimension reduction, Composite index, Multivariate analysis, eSS-CIDI
approach, Sustainable Society Index
1. Introduction
In the recent years composite indicators have become a valuable source
of information for policy makers, decision makers and the wider public (Greco
et al., 2018; Saisana et al., 2011). The OECD (2013) defines composite indicators
as metrics “formed when individual indicators are compiled into a single index,
on the basis of an underlying model of the multi-dimensional concept that is
being measured”. From this definition various questions arise (Nardo et al.,
2005) such as which indicators to include in the framework, whether to
normalize the data or no, how to decide on the importance of individual
indicators, and so on.
An important obstacle of composite indicators is that they usually aim to
measure a multidimensional phenomenon which cannot be measured with a
sole indicator (Decancq & Lugo, 2013). Therefore, the issue arises how to
156 | I S I W S C 2 0 1 9