近年来全球金融危机及欧债危机的爆发给世界经济带来巨大动荡的事实表明，人们对金融过程的认识、管理与控制能力还十分有限。正如知名学者Bouchaud（Nature，2008）所指出的，传统经济学理论不能理解可能发生大动荡的市场，因而该理论需要进行一场科学革命，需要从实际数据出发挖掘市场的特性。为此，本项目从微观的视点研究金融市场价格、市场过剩需求和流动性之间的动态关系，开发能描述该关系特性的、具有随机时变方差、含随机跳跃、长记忆性、杠杆效应、尖峰厚尾分布、且动态特性时变的复杂金融时间序列的一般化结构性建模方法。基于自组织状态空间建模、粒子滤波及并行计算等方法，开发针对参数时变、高阶的非线性非高斯型一般化金融市场微结构模型的状态与参数估计、随机跳跃检出的实用化估计方法。设计能有效利用该微结构模型的最佳投资策略。为分析金融市场变化、规避投资风险提供一套系统性的、有效的、实用化的解析方法及设计方法。

It has been verified that people' abilities to understand, manage and control financial processes are quite limited, because of the fact that the eruptions of the global financial crisis and the Europe debts crisis in recent years have been bringing a big havoc to the world global economics. Jean-Philippe Bouchaud (2008, in Nature) pointed out that 'classical economics has no framework through which to understand 'wild' markets', 'Economics needs a scientific revolution', and 'The prerequisites for more stability in the long run are the development of a more pragmatic and realistic representation of what is going on in financial markets, and to focus on data, …'. Therefore, this program researches the dynamic relationship between price and excess demand as well as liquidity of markets from micro-viewpoint, and develops a generalized and structured modeling method to the complex financial time series with stochastic time-varying variance, random jumps, long-memory, leverage effect, non-Gaussian distribution and time-varying dynamics. On the basis of self-organizing state-apace modeling, particle filtering and parallel computing techniques, the practical estimation method to state, parameters and detection of jumps will be proposed for the time-varying parameters and high-order nonlinear non-Gauss type generalized financial market microstructure model being studied. The optimal investment strategy is also designed based on the microstructure model. The aim is to provide a systematic, useful and practical analytic and design method to analyze financial markets' change and to avoid investment risks.