An Integrated Approach to Stock Selection and Ranking: Combining Shannon Entropy Technique, DEA, and Inverse DEA

This study proposes a novel, integrated framework that combines Shannon Entropy Technique (SET), Data Envelopment Analysis (DEA), and Inverse DEA (IDEA) for the objective evaluation and ranking of equities, with a case application to the tourism sector of the Taiwan Stock Exchange (TWSE). SET is employed for data-driven variable selection to reduce dimensionality across five financial perspectives, enhancing the discriminatory power of DEA. DEA identifies efficient stocks, while IDEA ranks them based on adaptability to output perturbations, providing a dynamic and interpretable view of efficiency. Empirical analysis reveals that only five out of sixteen stocks achieved full efficiency. The IDEA-based ranking uncovered distinct adaptability profiles, with some firms requiring minimal input changes to accommodate output growth, indicating operational flexibility. Sensitivity analysis and model comparisons with super-efficiency and cross-efficiency DEA confirm the robustness and consistency of the proposed method, even under adverse scenarios. This work advances financial decision-making by offering a reproducible, bias-minimized, and scenario-sensitive ranking model. Its transparent structure makes it especially valuable for investors, policymakers, and portfolio managers navigating volatile or crisis-prone markets.