Modeling social relationship strength in HR systems: a DEMATEL-ANP approach for the railway sector

This study develops a data-driven framework to evaluate employee social relationship strength using structured human resource (HR) records in the railway sector. We construct a hierarchical indicator system capturing three relationship bases—Geopolitical Relations, Blood Relations, and Industry Relations—and apply DEMATEL to model causal interdependencies among indicators, followed by ANP to derive their relative importance and generate hybrid weights. Using a railway HR database (2021–2023) as the empirical setting, the results indicate that kinship-related ties and workplace/education-based ties exert the strongest influence on relationship strength. Distinct from survey-based or communication-log-based approaches, the proposed framework operationalizes relationship indicators directly from administrative HR records, enabling scalable and auditable assessment in a highly regulated, safety-critical railway context. The proposed approach provides an objective and scalable tool for HR governance, supporting the identification of relational risks (e.g. favoritism and conflicts of interest) through interpretable, structured-data-driven measurement. By bridging social relationship theory with decision-science methods, this study offers operational insights for HR management in large-scale transport organizations.