中国股市尾部相关的渐进变化特征研究

发布时间：2018-04-30 11:14

本文选题：copula函数 + 开区间条件概率期望函数　；参考：《北方工业大学》2013年硕士论文

【摘要】：关于尾部分析的研究已经有了十分成熟的分析,copula函数的应用具体有很多的分类,从尾部相关系数的建立到上尾或下尾相关系数,再或是copula函数的选择,Kendall秩相关系数,等等。本文受到前人的提点和启发,是从尾部相关系数的概念出发,对于条件概率的定义以及根据不同copula函数捕捉不同的尾部特征,得出最佳的描述函数。当然也涉及了很多的细节,单独使用图形方法来决定最优Copula结果不是很明显,这里使用拟合优度检验来检查结果。然后,构造出不同区间的相关系数公式。之所以分为开区间和闭区间,是出于区间的不同情况考虑的。基于Copula的开区间条件概率期望函数区间是尾部的平均值,闭区间条件概率期望函数是某个区间的平均值。得到的结论能够给投资人一些参考,而且对于板块间的实证分析能够让投资人把握大的方向。
[Abstract]:The research on tail analysis has been very mature in the application of copula function. There are many kinds of classification, from the establishment of tail correlation coefficient to the upper tail or lower tail correlation coefficient, or the choice of copula function to Kendall rank correlation coefficient, and so on. Based on the concept of tail correlation coefficient, the definition of conditional probability and the capture of different tail characteristics according to different copula functions, the best description function is obtained. Of course, there are a lot of details involved. It is not obvious to use graphic method to determine the optimal Copula results alone. Here, we use the goodness of fit test to check the results. Then, the correlation coefficient formula of different intervals is constructed. The reason why it is divided into open interval and closed interval is due to the different conditions of the interval. The open interval conditional probability expectation function based on Copula is the mean value of the tail, and the closed interval conditional probability expectation function is the average value of a certain interval. The conclusions can give investors some reference, and for the empirical analysis between the plates can let investors grasp the big direction.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;O212.1

【相似文献】