人类通信模式中基于时间统计的实证研究  

Empirical Analysis on Temporal Statistics of Human Correspondence Patterns

作　　者：李楠楠[1] 张宁[1] 周涛[2,3] 

机构地区：[1]上海理工大学管理学院系统工程研究所,上海200093 [2]中国科学技术大学近代物理系,合肥230026 [3]弗里堡大学物理系

出　　处：《复杂系统与复杂性科学》2008年第3期43-47,共5页Complex Systems and Complexity Science

基　　金：上海市重点学科建设资助项目(T0502);上海市自然科学基金(06ZR14144)

摘　　要：最近关于人类通信模式的实证研究显示,信件的回复时间分布具有近似幂律的尾部,其中幂指数在电子邮件和普通信件中并不相同,前者为1.0,后者为1.5。统计了钱学森和鲁迅先生的信件数据,同样发现了近似幂律的回复时间分布,但是钱学森的分布指数是2.1,该发现暗示,将人类动力学划分为两个普适类的猜测可能是不充分的;信件发送的时间间隔分布也服从相同幂指数的幂律分布。同时,截取钱学森连续10年的信件得到的等待时间分布和间隔时间分布与总体基本相似,且幂指数也相同。这种巧合有可能暗示了潜藏在人类通讯模式中的某种未知机制。Recently, extensive empirical studies in human correspondence patterns show that the response times is heavy tailed - similar to a power law. The exponent is different in the e-mail and surface letters, which is 1.0 and 1.5 respectively. In this paper, through the statistical data from Qian Xuesen and Lu Xun＇s correspondence, we find the interval time distribution is a power-law with the exponent γ = 2.1 and T = 1.5 respectively. Further, the interevent time distribution which intercept from Qian xuesen＇ s letters for ten consecutive years, is also exponential distribution according with V zquez and others＇ research. Moreover, we find the response time distribution also follows the power-law with the same exponent about the interevent time, which may imply some unknown mechanism hidden in the patterns of human communication.

关 键 词：人类动力学 通信模式 幂律分布 

分 类 号：N93[自然科学总论] N94