Answers are grounded in this paper and cite their sources.
Criterion weights shape rankings in multi-criteria decision-making (MCDM), yet objective methods (Entropy, CRITIC, MEREC and SRCBECM) define importance through dispersion, pairwise correlation, or removal effects. This study proposes Spearman Correlation and Importance Vector-based Weighting (SCIV-Weight), which assigns weights in proportion to the absolute Spearman correlation between each criterion and its leave-one-out importance vector (LOO-IV), representing the shared structure of the remaining criteria. Formal properties include well-definedness, invariance to criterion order and positive affine transformations, and preservation of a criterion’s own correlation under a strictly increasing transformation of that criterion. It was evaluated on a 6 × 4 example, a 30 × 4 simulation, and ten scenarios covering dependence structures, sample sizes, outliers, ties and skewed margins. Under local perturbation, SCIV-Weight was more stable than Entropy and SRCBECM and close to CRITIC and MEREC; across independent samples, it was less stable than CRITIC and MEREC, and its weights may reflect noise when criteria share little common structure. In the Human Development Index data, SCIV-Weight-based TOPSIS rankings had the highest Spearman correlation with the official ranking, marginally above equal weights (0.9932 vs. 0.9930; 0.9982 vs. 0.9974 with log income). The log transformation changed SCIV-Weight weights by at most 0.0017, compared with 0.0174–0.4198 for the others.
No takes yet. Share an insight, caveat, or question.
Serkan Akoğul (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: