This paper proposes a novel hybrid framework for resilient supplier selection within the retail industry by integrating data mining approaches with MCDM techniques. Specifically, we present a methodological structure that effectively avoids rank reversal in supplier ordering, which has been a common issue in traditional methods. Our proposed hybrid approach comprises: (1) identify key resilience criteria for supplier selection in retail companies; (2) the Analytic Network Process (ANP) to weight these criteria and factor analysis to identify strategic resilient factors; (3) a K-medoids clustering technique to enhance the efficiency of the MCDM process; and, (4) the VIKOR method to rank the clusters. A case study involving a major retail chain is utilized to demonstrate the applicability of the proposed approach. The performance analysis is conducted in two parts. First, the results are compared with those from an ANP-VIKOR method for validation and verification, yielding a correlation coefficient of 0.856. Second, we introduce a procedure to assess the proposed method's ability to avoid rank reversal, resulting in a score of 0.83. Numerical results reveal that the proposed framework effectively addresses the complexities of resilient supplier selection in the retail industry, overcoming the limitations of traditional MCDM methods by preventing rank reversal.