摘要：鉴于已有的主成分分析法在效率改进方面体现出的不足，提出了一种改进的主成分分析法。该方法用于定量地选取评价指标，避免了主观选取的不确定性。通过输入与输出向量投影实现了对非DEA有效的决策单元的效率改进。给出了应用该方法的具体步骤，通过对8个港口进行指标选取和效率得分的计算与排序，验证了该方法的可行性并体现其相对于已有方法的优势。

Abstract： A new approach using Principal Component Analysis (PCA) on the efficiency improvement is proposed in order to improve the disadvantage of existing approach. Through the proposed method, the uncertainty of the subjective selection can be avoided by selecting indicators quantitatively, and the performance of non-efficient decision-making units can also be improved by projecting input and output vectors to the efficient frontier. Specific procedures of the improved method are given. Besides, an example of eight Chinese ports is presented here to verify the feasibility of the improved approach and its remarkable advantages over the existing one.