tsair-14-Most of the classical time series analyses require th.ppt

下载文档

5
0
约7.24千字
约 1页
2018-03-23 发布于甘肃
举报
版权申诉
保障服务

tsair-14-Most of the classical time series analyses require th.ppt

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。

* Most of the classical time series analyses require the time series to be stationary and/or linear. However, financial time series are usually nonlinear and nonstationary. In this study, we use the Hilbert-Huang Transformation HHT to investigate the dairy return of several international stock markets from 2005 to 2010. The HHT approach mainly consists of application of the empirical mode decomposition EMD and the instantaneous frequency analysis. We find that there is an obvious change in the behavior of the trading activity among these stock markets since the U.S. sub-prime mortgage credit crunch in 2008. Keywords: instantaneous frequency, empirical mode decomposition, HHT. Abstract In the stock time series analysis, the researchers usually use the logarithmic return scale for application. The logarithmic return is simply defined by ,where stands for the stock price at time t. Fig. 2 gives the logarithmic return of NASDAQ index and the first 3 IMFs obtained through the EMD method result in decomposition of 9 components and 1 residue. Therefore, the further application of the phase statistics approach is based on analysis on the first IMF. Consequently, we take an illustration with the NASDAQ index, the probability density functions of the amplitude for its first-IMF is roughly Boltzmann distributed. Beside this, we find from Fig. 3 that except for the first IMF, the phases of the other IMFs have almost equal probability for To investigate the correlative behaviors between any two different stock markets, we pair-wisely calculate the instantaneous phases between two indexes. Here, the phase differences is defined as Fig. 4 gives the probability distributions of phase differences between the first-IMFs of returns for two indexes for 2005–2010. We summarize some finding based various market regions and certain periods. 1 The third row plots provide the angle difference distribution of SP against the other indices. It can be seen, when subtracting against