Ranking the importance of multi-criteria evaluation indicators based on their transitivity

Abstract

In recent decades, multi-criteria methods have been increasingly used for the quantitative assessment of the development of socioeconomic systems. Their essence lies in weighted indicators, i.e., combining the values multiplied by the weights into one summarizing index. However, determining the significance of indicators is important in such approaches. It can be done in one or two stages. In the first case, the significance is assessed immediately, in the second case, the importance ranks of the indicators are determined before the assessment. Today, most people are satisfied with the first method, i.e., determining significance without knowing the importance ranks. This makes sense when the number of indicators is small. Socio-economic phenomena are, by their nature, complex and multifaceted, so in practice they manifest in many aspects. Therefore, their condition can be adequately assessed only with a large number of indicators. The significance of the indicators of such systems is assessed by comparing the importance of paired indicators. However, in the presence of a large number of indicators, there are constantly recurring problems - excessive volumes of expert evaluations and, as a result, a decrease in the adequacy of the evaluation. Transitive analysis of index importance (TAII) is the proposed methodology that allows to significantly increase the number of evaluated indicators while reducing the volume of expert evaluations and increasing their adequacy. This can be achieved by integrating their transitivity as a property into the ranking procedure of determining the importance of indicators. In this way, the volume of expert evaluations can be reduced by 40%. The suitability of the proposed methodology has been verified using real problems.