基于PCA与MLE方法的人群分类新方法研究  

A Novel Method for Classification of Crowd Based on PCA and MLE Methods

作　　者：张成[1] 刘亚东[1] 谢彦红[1] 李金娜[1] 张可意 李元[1] 

机构地区：[1]沈阳化工大学数理系,辽宁沈阳110142 [2]大连理工大学公共管理与法学院,辽宁大连116024

出　　处：《沈阳化工大学学报》2015年第2期168-171,共4页Journal of Shenyang University of Chemical Technology

基　　金：国家自然科学基金资助项目(60774070;61174119);国家自然科学基金重点课题资助项目(61034006)

摘　　要：针对人群分类管理问题,提出一种基于主元分析方法的新的单变量贡献度方法.首先,研究各行为属性对主元的贡献率,确定人群关键行为属性.然后,在确定关键属性的基础上,利用最大似然估计方法对人群进行分类管理.应用人群分类管理方法到警方侦破案件中,可大大减小案件的搜索范围,达到提高警方办案效率的目的.通过实例验证了方法的有效性.A novel classification of the crowd is investigated based on principal component analysis method in this paper. First,the contribution rates provided by each behavior attributes for principal component are studied to determine the key behavioral attributes. Then,based on the identification of key attributes,the residents are divided into several groups by the maximum likelihood estimator method. The presented classification method of the crowd not only narrows the search areas,but also improves the efficiency of the police investigations. Finally,a case is given to showthe effectiveness of the proposed method.

关 键 词：主元分析(PCA) 属性分类 最大似然估计(MLE) 

分 类 号：C939[经济管理—管理学]