一种演绎金融数据沙堆演化过程的新方法  

A New Method to Interpret Financial Data Sandpile Evolution Process

作　　者：郭兆纲[1] 魏宝社[2] 胡振鹏[1] 

机构地区：[1]南昌大学管理学院 [2]东北财经大学社会与行为跨学科研究中心

出　　处：《数量经济技术经济研究》2016年第5期122-142,共21页Journal of Quantitative & Technological Economics

摘　　要：用递归图方法演绎金融数据的沙堆演化过程是我们为了研究金融数据中所蕴含的沙崩和跃升现象而提出的一种创新方法。该方法基于重构相空间理论,通过对比分析递归图中类分形自相似结构的演化过程与沙堆堆积过程之间的相似性,从直观分析沙堆堆积、滑落及坍塌的特征入手,构建类分形自相似结构的沙堆演化模型,提出了驻留比率、滑落比率及时间间隔的边界点等概念工具,研究金融数据中所蕴含的沙崩和跃升现象。本文重点介绍该方法提出的思路和模型的构建过程,并以上证综合指数为例验证了方法的有效性。Using recurrence plot to interpret financial data sandpile evolution process is an innovative method to study the sand avalanche and zooming phenomena contained in financial data. Based on the phase-space reconstruction theory and the comparison of the similarities between evolution process of fractal-like self-similar structure and sandpile heaping process in the recurrence plot, this paper starts with intuitive analysis of sandpile heaping, slide and sand avalanche to construct a sandpile evolution model with a fractal-like self-simi- lar structure and puts forward such general tools as lingering ratio, slide ratio, boundary points of time intervals to study the sand avalanche and zooming phenomena contained in fi- nancial data. This article mainly introduces the train of thought and construction process of the model proposed in this method and verifies the effectiveness of this method by taking the Shanghai composite index as an example.

关 键 词：递归图 类分形自相似结构 沙堆模型 沙崩 跃升 

分 类 号：F224.9[经济管理—国民经济] F831.5