基于矢量空间的群体聚类方法研究  

Research on the group clustering method based on vector space

作　　者：徐选华[1] 陈晓红[1] 

机构地区：[1]中南大学商学院,湖南长沙410083

出　　处：《系统工程与电子技术》2005年第6期1034-1037,共4页Systems Engineering and Electronics

基　　金：国家自然科学基金委国家杰出青年科学基金(70125002);国家教育部博士点基金(2000053306)资助课题

摘　　要：通过对群体成员的思维模式进行分析,基于矢量空间提出了群体中任意两个成员偏好矢量相关性度量的一般范数以及基于该矢量相关度的启发式算法,通过执行该使算法使群体中的成员形成若干个不同的聚集。在此基础上,定义并计算群体中各个聚集的偏好和整个群体的偏好,定义并计算群体中各个聚集的一致性指标和整个群体的一致性指标,为群体决策提供依据。最后通过一个实例对算法和一致性指标等进行验证。By analyzing the thinking pattern of group members, the general norm of relativity measurement of preference vectors of any two members in the group and the heuristic algorithm based on this vector relativity degree are put forward based on vector space. With the algorithm, several different clusters are formed in the group. On this basis, the preferences of every cluster of the group and the preference of the entire group, as well as the coherence indexes of every cluster of the group and the coherence index of the entire group are defined and calculated, which provides the basis for group decision. Finally, the algorithm and coherence indexes proposed are validated by an example.

关 键 词：矢量空间 群体聚类 群体一致性 群体决策 

分 类 号：C934[经济管理—管理学]