关键词: 个性化语义, 共识, 可信度风险, 专家权重, 柔性语言表达

Abstract: Current flexible linguistic group decision-making methods still exhibit shortcomings in personalized individual semantic (PIS) modeling, adaptive consensus mechanisms, and the consideration of experts' credibility risks, which constrain the efficiency and reliability of consensus reaching process. To address these issues, this paper proposes an adaptive flexible linguistic consensus mechanism that incorporates experts' credibility risks, shifting from traditional segmented optimization to a dynamically coupled systemic coordination. The main contributions are: (1) A PIS modeling method based on subjective-objective cross‑validation is proposed, which directly optimizes the semantics of flexible linguistic expressions by minimizing the deviation between subjective and objective preferences while maximizing their consistency, thereby avoiding information loss and enhancing the overall preference consistency. (2) A dynamic credibility risk evaluation and adaptive weight allocation mechanism based on ranking consistency deviation is designed, which realizes real‑time coupling between risk perception and weights of experts, effectively suppressing the interference of high‑risk experts on consensus. (3) An adaptive consensus feedback mechanism integrating credibility risk tracking is constructed, which simultaneously updates preferences, risks, and weights during consensus iterations, thereby achieving a coordination improvement in consensus efficiency and decision robustness. Case studies, sensitivity analyses, and comparative experiments verify the effectiveness and reliability of the proposed method.

Key words: personalized individual semantic, consensus, credibility risk, expert weight, flexible linguistic expression