基于业绩持续性的证券投资基金聚类与实证研究  

Study on the Clustering of Security Funds Based on the Performance Persistence and its Demonstration

作　　者：时希杰[1] 吴育华[1] 李亚瑞 

机构地区：[1]天津大学管理学院,9013天津300072 [2]汉唐证券研究所,深圳518053

出　　处：《数理统计与管理》2005年第4期66-69,117,共5页Journal of Applied Statistics and Management

摘　　要：目前基金的分类方法大多是从基金的性质特点出发的主观分类方法,没有反映出基金资产的实际运作效果。本文采用聚类的思想,以基金的业绩表现为基础,从业绩持续性的角度提出了一种新的分类方法:基于业绩持续性的基金聚类。通过对基金业绩持续性的研究,构造了一个业绩持续指数,并用该指数对样本进行聚类。实证研究结果表明,该分类方法是可行和有效的。Most current classification methods of security funds are based on the characters of the fund itself, which are subjective rather than objective. It cannot reflect the actual operation of assets. In this paper, clustering of security funds based on the performance persistence, a new classification method is presented from the view of history performance, using the thought of clustering. A performance persistence index is constructed when studying the performance persistence of funds. Clustering based on this index is also discussed. The result of demonstration on history data shows that this method is feasible and effective.

关 键 词：证券投资基金 分类 H—L方法 业绩持续指数 聚类 

分 类 号：F830.91[经济管理—金融学]