A comparative fuzzy MCDM framework for employee evaluation: Fuzzy TOPSIS versus Fuzzy AHP–TOPSIS on the IBM HR analytics dataset
This study compares equal-weight Fuzzy TOPSIS with hybrid Fuzzy AHP–TOPSIS for multidimensional employee evaluation. Both models use the synthetic IBM HR Analytics dataset (1,470 employees) and the same seven criteria; only the weighting rule changes. Expert-derived weights were highest for Job Involvement (0.241140), Years in Current Role (0.200279), and Training Times Last Year (0.169016). The rankings were strongly associated (Spearman 0.904389; Kendall’s tau-b 0.738746), yet individual positions moved substantially (mean absolute shift 139.169; maximum 601.5). Top-10, Top-20, and Top-50 overlaps were 60%, 55%, and 66%. Fourteen employees shifted by more than 500 positions. Sensitivity analysis of both models, covering 56 weight perturbations, kept within-model Spearman correlations above 0.990. An alternative coding that preserves the numerical detail of two criteria left the global weighting contrast intact (Spearman 0.905428) while changing Top-10 membership. A secondary vertex-distance Fuzzy VIKOR ranking that used the same expert weights agreed more closely with hybrid TOPSIS (Spearman 0.898969) than with equal-weight TOPSIS (Spearman 0.718161). Because the dataset is synthetic, the rankings are a reproducible decision-support benchmark, not evidence about a real workforce.
Authors
- Zarifa Gasim Jabrayilova (ORCID: https://orcid.org/0000-0002-9661-5805)
- Fuad Hüseynli
Institutions
- Azerbaijan Technical University (AZ)
- Institute of Information Technology (AZ)
Publication Details
- Journal
- Journal of Electrical Systems and Information Technology
- Published
- 2026-11-18
- DOI
- https://doi.org/10.1186/s43067-026-00420-3
- Primary Topic
- Multi-Criteria Decision Making
- Type
- article
- Field-Weighted Citation Impact
- 0.00