证券市场高频数据双相行为的仿真  

Numerical study on the two-phase behavior of financial markets

作　　者：刘兴华[1] 梁晓蓓[2] 杨海波[1] 

机构地区：[1]山东经济学院信息管理学院,济南250014 [2]同济大学经济与管理学院,上海200092

出　　处：《系统工程理论与实践》2009年第12期147-153,共7页Systems Engineering-Theory & Practice

基　　金：国家社会科学基金(08BJL020);教育部人文社科基金(07JC790066);上海市科技发展基金软科学研究项目(086921005);国家自然科学基金(70832005)

摘　　要：Plerou,et al以证券市场高频数据为研究对象,通过实证分析发现了证券市场的"双相行为"(Two-phase behavior offinancial markets).这个现象是个复杂性涌现,Plerou猜想引起这个现象的原因是私人信息.该文在少数派博弈模型的基础上,建立了一个市场人工模型,仿真得到了同样的特征,并发现智能代理多样化的归纳推理模式是产生"双相行为"的关键因素,市场微结构也有一定的影响,而私人信息不是关键因素.An empirical study on high-frequency data of financial markets which was published on Nature found a new stylized fact so called two-phase behavior.This phenomenon is a complex emergency,the author Plerou suggested this phenomenon may be caused by the diffusing of private information.In this paper,we developed a new artificial stock market based on the extended minority game.We simulated on this new model and reproduced the two-phase behavior.We also found that the key factor causing two-phase behavior is inductive reasoning which is main character of investor＇s behavior.Microstructure is another reason impacts two-phase behavior by combined working with inductive reasoning while the privative information is not the essential reason to produce two-phase behavior.

关 键 词：智能代理建模 人工证券市场 多样性和归纳推理 双相行为 

分 类 号：N941[自然科学总论—系统科学] F830[经济管理—金融学]