Document Type : Research Paper
Authors
Shahid Bahonar University of Kerman, Kerman, Iran
Abstract
Triangular fuzzy numbers (TFNs) are widely used in MCDM problems with linguistic expert judgments. A key challenge is determining appropriate upper and lower bounds, especially with few experts, as symmetric scales often yield near‑crisp values and biased results. This paper proposes a bootstrap‑based method to estimate TFN boundaries using confidence intervals from expert evaluations, enabling data‑driven, statistically grounded fuzzy triangles that better reflect uncertainty. The method is integrated into fuzzy DEMATEL. Simulations show it discriminates causal relationships more clearly than classical fuzzy DEMATEL. An empirical case study on inflation determinants confirms that the bootstrap‑based model produces results more consistent with economic behavior and expert expectations, demonstrating its effectiveness in improving the reliability of TFN‑based decision‑making.
Keywords
)