论文部分内容阅读
针对STSA方法在金融时间序列分析中的缺陷提出了运用EMD与STSA结合的改进方法。以上证指数、深证成指、建筑指数、金融指数、地产指数、上证商业6种指数的收益数据作为研究样本,利用EMD方法分解提取出一系列反映原始序列不同时间尺度信息的分量,通过对各分量进行STSA分析后发现导致原始序列多变化模式的原因。在此基础上提出了通过单一变化模式分量对原始序列变化趋势进行估计的条件和限定范围,实验结果表明,该方法在提取和分析时间序列变化模式方面具有独特的优势,具有较高的预测精度和实用性。
In view of the shortcomings of STSA method in financial time series analysis, an improved method using EMD and STSA is proposed. Using the income data of Shanghai Stock Index, Shenzhen Stock Exchange Index, Shenzhen Construction Index, Financial Index, Real Estate Index and Shanghai Stock Exchange as the research samples, the EMD method is used to extract a series of components that reflect the information of different timescales of the original sequence. STSA analysis of each component found the reason for the multiple sequence pattern of the original sequence. On the basis of this, the conditions and limits of estimating the trend of the original sequence through a single change pattern component are proposed. The experimental results show that this method has unique advantages in extracting and analyzing the time series variation pattern and has higher prediction accuracy And practicality.