论文部分内容阅读
本文尝试提出一种全新的证券分类方法:即运用Copula-GARCH类模型来计算两两证券价格序列间的静态Copula系数、以测算其同期相依性,并据此构建各样本间的相似矩阵,再根据相似矩阵对相关证券进行模糊聚类分析,从而对相关证券进行分类。本文所提出的证券分类方法不以有效市场假说为基础,比默认有效市场假说的证券分类方法更具普适性,更贴近于证券市场的实际情况、能更真实地反映证券市场的内在机理和波动特征。本文的实证研究,量化测定了中证100指数成份股的同期相依性,并据此对中证100指数的成份股进行分类,印证了本文所提出的基于Copula-GARCH类模型的证券分类方法的有效性和实际价值。
This paper attempts to propose a new method of securities classification: the use of Copula-GARCH class model to calculate the static Copula coefficient between two securities price series to measure the same period dependencies, and then build a similar matrix between the samples, and then According to similarity matrix, the related securities are classified by fuzzy cluster analysis. The method of securities classification proposed in this paper is not based on the effective market hypothesis and is more general than the securities classification method of the default effective market hypothesis. It is more close to the actual situation of the securities market and can more accurately reflect the intrinsic mechanism of the securities market. Fluctuation characteristics. The empirical study of this paper quantifies the contemporaneous dependence of the constituents of the CSI 100 Index and classifies the constituents of the CSI 100 Index, confirming the proposed securities classification method based on the Copula-GARCH class model Validity and real value.