A Visualization Analytics Framework for Interpreting Fuzzy Inference Systems in Supply Chain Management

Fuzzy inference systems (FISs) have been widely applied to support decision-making and operational functionalities in supply chain management (SCM), leveraging highly interpretable fuzzy inference rules (FIRs) expressed through linguistic terms to enable transparent and flexible nonlinear modeling, planning, control, and optimization. However, in many practical settings, multiple FIRs may be activated simultaneously with varying degrees of relevance, creating ambiguity for decision-makers, similar to interpretability challenges observed in ensemble models such as random forests. To address this issue, this study integrates explainable artificial intelligence (XAI) techniques into FIS-based SCM applications to enhance interpretability and user trust, employing visualization-driven methods such as partial dependence plots (PDP), SHAP analysis, and locally interpretable model-agnostic explanations (LIME) to provide insight into inference behavior. Building on these approaches, an FIS-based incremental interpretation framework is proposed to systematically reduce ambiguity and support human-centric decision-making. The proposed methodology is validated through a real-world supply chain case study, evaluated through expert-based assessment, and further examined using statistical hypothesis testing to establish its effectiveness and practical relevance.