经济时间序列相空间重构与混沌特性判定研究

经济时间序列相空间重构与混沌特性判定研究 Theory and Practice on Reconstructed Phase Space of the Economic Time Series and Chaotic Characteristics 郁俊莉 YU Jun-li~1 1 2 3 first-author 王其文 WANG Qi-wen~1 1 2 3 韩文秀 HAN Wen-xiu~2 1 2 3 北京大学光华管理学院 北京大学光华管理学院 School of Guanghua Management, Peking University School of Guanghua Management, Peking University 天津大学管理学院,北京,100871 天津大学管理学院,北京,100871 School of Management,Tianjin University School of Management,Tianjin University ,天津,300072 ,天津,300072 提出通过相空间重构及Lyapunov指数来判定时间序列的混沌特性.给出了时间序列相空间重构中确定最小嵌入维数的伪邻点法及确定时滞参数的自相关函数法;提出了一种计算Lyapunov指数的实用方法;最后,以上证综合指数收益率时间序列为例研究了经济时间序列的混沌特性判定问题. The purpose of this article is to judge the chaotic characteristics of time series by using phase space reconstruction and Lyapunov exponents. We discuss the method of the false neighboring and the auto-relativity function for determining the time delay parameter and minimum embedding dimension of the reconstructed time series phase space. Further more, we also put forward a method for calculating the Lyapunov exponents. In the last, a case study on judging the chaotic characteristics with the logarithm yield of the Shanghai Stock Market is performed. 相空间重构 伪邻点法 自相关函数法 Lyapunov指数 phase space reconstruction false neighboring method auto-relativity function method Lyapunov exponents F224 中国博士后自然科学基金资助项目（2003033080） 2004-01-01 2021-04-01 1