Decision-making in complex environments increasingly depends on the ability to transform heterogeneous textual records into structured analytical evidence. However, existing studies on festival-related texts are predominantly qualitative and rarely provide operational decision models that convert textual evidence into measurable criteria for management and policy support. To address this gap, this study proposes a corpus-driven multi-criteria decision support framework that integrates corpus linguistics, functional text analysis, and a structured multi-criteria decision-making (MCDM) procedure. Using the Hmong Sister Festival in Guizhou, China, as a case study, a structured diachronic corpus database is constructed from historical documents and multi-source textual materials. Quantitative analytical techniques, including keyword extraction, co-occurrence network analysis, topic modeling, and principal component extraction, are employed to identify dominant textual structures and temporal patterns and to transform textual evidence into measurable decision criteria. The extracted indicators are then normalized, weighted through a structured criteria-weighting process. The results reveal that the textual system exhibits strong diachronic continuity while simultaneously reflecting distinct context-dependent characteristics. By linking corpus-driven text modeling with an explicit multi-criteria decision structure, the proposed framework provides a reproducible and computationally transparent approach for supporting evidence-based decisionmaking in complex decision environments, including cultural heritage management, tourism planning, knowledge management, documentation-intensive governance, and text-enabled decision analytics.

multi-criteria decision-making (MCDM), TOPSIS, corpus-driven text analytics, unstructured textual data