Abstract

The decision-making process in phone refurbishment has traditionally relied on subjective methods, such as the Analytic Hierarchy Process (AHP), which depends heavily on expert judgments. This paper introduces a data-driven approach to decision-making in cell phone refurbishment process, aimed at reducing subjectivity and increasing objectivity. By integrating Random Forest and SHAP (SHapley Additive exPlanations) importance values with linear regression, we perform an AHP style analysis based on data-derived weights rather than human judgments. This approach ensures more accurate prioritization in the decision-making process. In this paper, we focus on refurbishment cost, which is influenced by several factors. We conduct an AHP style analysis using these factors to assess their impact on refurbishment costs. Additionally, we calculate the entropy and consistency ratio (CR) to validate the robustness of our methodology. The results demonstrate that this approach not only improves decision-making but also provides a replicable framework for optimizing refurbishment processes, positively impacting costs, quality of refurbished phones, and turnover time.