Nonlinear Characteristics And Application Research Of Chinese Capital Market

In this context,the financial system is seen as a dissipative system with high dimension,high openness,far from equilibrium and entropy exchange instead of traditional simple linear system.Therefore,to study Chinese capital market with complex science has important theoretical and practical significance.The research helps to expand the application scope of industrial engineering(IE)on one hand,and have suggestions for China's developing financial market on the other hand.Based on previous studies,the research of this paper mainly includes following sections.(1)Time series in Chinese stock market is proved to have nonlinear characteristics with surrogate data method and BDS method.By determining the indicators such as delay time,embedding dimension,correlation dimension and the largest Lyapunov exponent,the market is proved chaotic,and therefore non-linear prediction is feasible.(2)The paper focuses on fractal characteristics in capital market.The stock market and commodity market have long-range correlation within a certain period of time and follow the Fractal Market Hypothesis(FMH).With MFDFA analysis,market is considered to be multi-fractal rather than single fractal.This feature comes from the interaction of internal market factors and external government factors.With MFDCCA analysis,we found multi-fractal cross-correlation between the two markets.Whether the cross-correlation is positive or negative depends on the order.Meanwhile,the cross-correlation between the markets is more pronounced than the autocorrelation in single market.Conclusions have important practical significance for both managers and investors.Managers should use scientific theories to grasp the operation laws of the market.For investors,combine fundamental analysis and technical analysis to grasp market trends and obtain excess profits.(3)Before non-linear prediction,firstly use wavelet neural network for noise reduction,then use RBF neural network and adaptive Volterra method to predict market movements.The results show that both methods have some predictive capability,and multi-step Volterra method is better than RBF method when dealing with this prediction.(4)Consider factors affecting the evolution of the financial ecosystem from the macro view,and build system dynamics model.For the driving factors,there are three factors including technological progress and institutional innovation,market demand,competition in the institutions.For supporting conditions,it is divided into four aspects including economic base,government behavior,social credit situation,and intermediary service level.Therefore,the government should coordinate economic development and create good competitive environment for financial institutions to improve products and services to meet market demand.