金融市场联动形态结构的非线性分析

Nonlinear analysis on the pattern structures of connection between financial markets

从微观层面分析货币市场与资本市场的联结问题,通过构建支持向量机(SVM)和Copula函数的集成系统,研究金融市场联结途径与形态结构,深层次挖掘两个市场互动的规律.针对金融时间序列的非线性、非平稳特性,利用改进的样本加权支持向量机估计SHIBOR和股票市场价格指数收益率的边缘概率分布函数.然后,再采用优选的Copula函数进一步分析它们的联合概率分布及其在不同情形下的变化情况,获得了货币市场与资本市场联动的相依结构与非线性联结的动态特性.实证分析表明,1Y-SHIBOR和股票市场价格指数收益率的联合分布概率在两者不同变化方向和幅度上表现出不同的变化规律,并存在不对称性.压力测试分析也呈现出相似的结果.

This paper analyzes the connection between the money market and the capital market in a micro level sense,investigates the connection approaches and their pattern structures,and mines deeply the dynam-ics of interactions between the two markets through the integration of support vector machines（ SVM）into cop-ula functions. The modified sample-weighted SVM is employed to estimate the marginal probability distribution functions of the money market price,SHIBOR,and the stock price index return of the capital market in terms of the nonlinearity and instability of financial variable time series,and the optimized copula function is then used to analyze the joint probability distributions of the two financial variables and their change scenarios under various circumstances,with the dependence structures between the two markets and the dynamics of the non-linear connection obtained. The empirical analysis shows that the joint distribution probability of 1Y-SHIBOR and the stock price index return exhibit different structure characteristics under different directions and magni-tudes when changing the two variables above,there exists asymmetries,and the stress test analysis gives simi-lar results.