An Explainable Artificial Intelligence Approach for Multi-Criteria ABC Item Classification

Multi-criteria ABC classification is a useful model for automatic inventory management and optimization. This enables rapid of items into three groups, having varying managerial levels. Several methods, based on different criteria principles, were proposed to build the classes. However, existing methods operate as black-box AI processes that only provide assignments classes without providing further explanations. The multi-criteria nature problem makes utilization interpretation item difficult, information. Decision makers usually need additional information regarding important characteristics crucial in determining because such can help managers better understand groups make decisions more transparent. To address this issue, we propose two-phased explainable approach eXplainable Artificial Intelligence (XAI) capabilities. provides both local global explanations built at class levels, respectively. Application firm, specialized retail sales, demonstrated its effectiveness generating accurate interpretable Assignments well-explained item’s criteria. results particular application have shown significant impact profit, customer priority had an