基于Copula函数的汽车与钢铁股票价格相关性分析

Correlation Analysis of Automobile and Steel Stocks Prices Based on Copula Function

本文基于Copula函数理论,研究了我国汽车与钢铁股票价格之间的相关性。在边缘分布的确定上,采取非参数核密度法估算汽车与钢铁股票的边缘分布函数;在Copula函数的参数估计上,通过极大似然法对引入的五种不同Copula函数进行参数估计;在函数模型的选择上采用欧式平方距离法则和AIC、OLS最小法则相结合的方法进行优选。研究表明t-Copula函数模型能够更好地描绘汽车–钢铁之间的相关结构,两者存在中等程度的正相关性,且上下尾相关系数相等。

The study is based on Copula function theory and investigates the correlation between automobile and steel stock prices in China. Determining the marginal distribution using non-parametric kernel density estimation for automobile and steel stock returns;Estimating parameters for five different Copula functions introduced through maximum likelihood estimation;Selecting the functional model using a combination of Euclidean distance method and AIC, OLS minimization method. The research indicates that the t-Copula function model can better depict the correlation structure between automobiles and steel, showing a moderate level of positive correlation between the two, with equal upper and lower tail correlation coefficients.