中文摘要：

有效挖掘群决策信息，并为管理者提供良好的决策支持一直是理论界所关注的课题。本文基于竞优思想，针对层次结构系统，依托目标规划模型和聚类分析方法，构建了群体竞优行为特征的群识别方法。本方法从最有利于体现被评价群体成绩的角度，运用目标规划模型，从每一组调研数据中挖掘被评价群体的优势特征，再利用聚类分析方法，从“优势特征集”中识别群体竞优特征的分布状况，进而可以清晰而具体地了解被评价群体的竞优行为特征。通过逻辑推导和实际算例，分别论证了本方法的有效性和可行性。

英文摘要：

It is an universally concerned topic in decision-making theoretical field about research on effective information mining which provides support for the managers. Based on the thought of match better confirming to the natural rule, a group identification method of group behavior recognition confirming to the natural rule is built up relying on goal programming model and cluster analysis method according to hierarchical structure. From the view of its being the most beneficial to the group evaluated, by means of goal programming model, the advantage character according to each group of datum is analyzed, and the distribution state of match better conforming to the natural rule is obtained by cluster analysis of the set of advantage character, The evaluated group behaviour gesture is obtained. The effectiveness and feasibility of this method is proved by logical deduction and practical calculation example.