基于选举模型理论研究股市特性  

Features analysis of stock markets based on Voter model theory

作　　者：牛红丽[1] 王军[1] 

机构地区：[1]北京交通大学理学院,北京100044

出　　处：《北京交通大学学报》2012年第3期138-144,共7页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家自然科学基金资助项目(70771006;10971010)

摘　　要：基于交互粒子系统之一的选举模型理论,构造了一个股票价格方程(模型)来实现股市价格和收益的模拟.文中讨论了选举模型理论中3个重要参数,强度、初始密度和网格维数,对股票收益统计特性及幂律分布的影响.为验证价格模型的有效性,对比上证综指、香港恒生指数及模拟收益,通过自相关系数分析、经典R/S分析法和修正R/S分析法来研究以上3个时间序列的长期依赖性.同时计算并检验了著名的Hurst指数,求取了以上3个金融序列的记忆周期.A random stock price model for simulation of stock prices and returns on the basis of Voter Model theory, which is one of the interacting particle systems, is constructed in this paper. The influences on statistical behaviors and power law distribution that three significant parameters in the Voter Model which are intensity, initial density and lattice dimension have been discussed and studied. Using different methods like autocorrelation coefficient analysis, classic R/S analysis and modified R/S analy- sis, the long-range dependence for SSECI, HSI, and simulative returns sequences for price model are investigated respectively to justify the effectiveness of the price model. Meanwhile, the significant Hurst exponent is given and its significance is tested while memory periods for three time series above are computed.

关 键 词：股票价格 选举模型 价格公式 模拟 统计特性 长程依赖性 

分 类 号：O211.6[理学—概率论与数理统计]