Abstract: Observing the security and stability of the real economy and social order through the signals transmitted by the financial market has attracted more and more attention. However, the current research in industry and academia mainly depends on economic theories and econometric models. Obtaining the required data is faced with the problems of low sample frequency and difficult statistics, as well as its interpretability is also limited. Therefore, this paper chooses to obtain financial data with low cost, strong timeliness, authority and reliability, and regards the financial time series as a financial signal hiding the function information of social and economic system. The Hilbert Huang transform（HHT） method composed of empirical mode decomposition（EMD） and Hilbert spectrum analysis（HAS） is selected to analyze the financial signal, capture the emergencies that impact the economy from the financial signal through EMD method, and measure the impact of emergencies on the society through HAS method. The experimental results show that HHT can overcome the problem that it is difficult to mine practical information due to the aliasing effect of financial signals, and can quickly identify and warn the actual emergencies.

Key words: signal analysis, spectrum analysis, Hilbert Huang transform（HHT）, empirical mode decomposition（EMD）

摘要： 通过金融市场传递的信号观察实体经济和社会秩序的安全和稳定受到了越来越多的关注。然而当前产学界的研究较多依赖于经济学理论和计量模型，获得需要的数据面临样本频次低、统计难度大的问题，其可解释性也受到了限制。为此，该文选择获得成本低、时效性强、权威可靠的金融数据，将金融时间序列视为隐藏社会和经济系统运行信息的金融信号。选择由经验模态分解（EMD）和希尔伯特谱分析（HAS）组成的希尔伯特-黄变换（HHT）方法对金融信号进行分析，通过EMD方法从金融信号中捕获对经济形成冲击的突发事件，通过HAS方法测度突发事件对于社会产生的影响程度。实验结果表明，HHT能够克服金融信号由于混叠多种效应难以挖掘具有实际意义的信息的问题，对于实际发生的突发事件能够做到快速识别和预警作用。

关键词: 信号分析, 频谱分析, 希尔伯特-黄变换, 经验模态分解