社会关系网络中人类行为标度特征实证分析.pdf

DOI:10.14018/j.cnki.cn13-1085/n.2013.08.104 ·298 · 价值工程 社会关系网络中人类行为标度特征实证分析 The Empirical Analysis of the Scale Feature of Human Dynamics in the Social Network 刘红玲① LIU Hong-ling ；郝进欣① HAO Jin-xin ；邓明森①② DENG Ming-sen （①贵州师范学院物理与电子科学学院，贵阳550018 ；②贵州师范学院应用物理研究所，贵阳550018） （①School of Physics and Electronic Sciences ，Guizhou Normal College ，Guiyang 550018，China； ②Institute of Applied Physics ，Guizhou Normal College ，Guiyang 550018，China） 摘要： 本论文利用社会关 系网络中某大型社 区人们之 间电子邮件通信数据库 ，对人类动力学标度特征进行实证分析研究 。在项 目合作组成 员验证 了社 区中个体发送 电子邮件的时间间隔统计特性满足幂律分布的基础上 ，分析 了社 区成 员接收、回复电子邮件的 时间间隔统计特性均具有明显偏 离指数分布的尖峰重尾 ，用幂函数能够很好的拟合 ，发现个体层 面的幂律指数并非传统的具有 离散 指数的普适类。 Abstract: In this thesis, the empirical analysis of the scale feature of human dynamics has been made through a database of email communications between a large community of people in the social network. The project cooperation members have proved the statistical properties of time interval distribution by individuals email sending in the community is corresponding to the power-law distribution. On the basis of which, the paper analyzes the statistical properties of time interval distribution by individuals email sending, receiving and reply in the community. The study obviously shows they all have the leptokurtosis and fat-tailed which deviate from the exponential distribution. We found that the power-law index of the individuals level is not a traditional universality class that has discrete index. 关键词： 人类动力学；标度特性 ；电子邮件 ；时间间隔分布 ；幂律分布 Key words: human dynamics ；scale feature ；email ；time interval distribution ；power-law distribution 中图分类号： 文