股票交易时间间隔分布特征的实证研究  

Empirical Research on the Stock Trading About the Distribution of Inter-Trade Duration

作　　者：郑木华[1] 姜罗罗[1,2] 赵明[1] 

机构地区：[1]广西师范大学物理科学与技术学院,广西桂林541004 [2]温州大学物理与电子信息工程学院,浙江温州325035

出　　处：《复杂系统与复杂性科学》2012年第2期31-35,共5页Complex Systems and Complexity Science

基　　金：国家自然科学基金(11047012;10805045;11165003);教育部重点项目(210166)

摘　　要：通过对中国股票市场中大量投资者的股票交易数据进行统计分析,发现个体买入和卖出股票的时间间隔分布具有幂律分布的特征,均能通过阈值为0.9的Kolmogorov-Smirnov统计性假设检验且幂指数几乎一致,可能体现着人类买卖这种相对应行为的相关性。股票交易次数和交易金额的分布也具有明显的胖尾现象,但并不具有幂律分布的特征。结果表明中国股市仍以小投资者居多,而且投资者的平均交易次数偏少。In this paper, abundant stock trading records of investors in Chinese stock market have been analyzed and the non-Poisson property is found in the distribution of inter-trade duration on the individual stock exchange. The power-law distribution has been found in the distribution of inter-trade duration about buying or selling the stocks. The Kolmogorov-Smirnov method with threshold 0.9 has been used to test whether the distribution obeys power-law. Or not the distri- butions of inter-trade duration of buying and selling behaviors have almost equal power exponent, which indicates the close relation between the human buying and selling behaviors. Furthermore, obvious heavy-tailed phenomenon has been found in the transaction frequency distribution and transaction amount distribution, but they are failed to fit the power-law distribution. The results also show that most investors hold small amount of stocks and trade stocks infrequently, which reflects the particular character of Chinese stock market.

关 键 词：人类动力学 非泊松特性 时间间隔 幂律分布 

分 类 号：N94[自然科学总论—系统科学]