【摘要】 对人类行为的定量分析是理解很多复杂社会经济现象的基础。基于BBS网络,给出了根贴的受欢迎度和用户的活跃性的实证研究,发现他们的分布可以用幂律函数来刻画且有胖尾特征,而相邻帖子的时间间隔概率分布近似满足延拓指数分布;用户的发贴时间间隔概率分布有一个明显的峰值,随着用户活跃程度的降低,概率分布变得越来越没有规律,胖尾特征越明显;此外,热点话题在典型用户所发的帖子中比例变化不大,这对发现论坛的热点话题有很大帮助。更多还原

【Abstract】 Quantitative understanding of human behavior supplies basic comprehension of the complexity phenomena of many socio-economic systems.Based on the log data of BBS,this paper presents empirical exploration of popularity of FileRoot and activity of users and finds that the distributions of both them have fat tail feature.It also finds that the distributions of interevent time between two consecutive files can be description approximatively by stretched exponential distribution,but the distributions of interevent time of major users have peak values and with the reduction of user activity the fat tail become more and more irregular indicating different habits for users with different activities.Besides,the change of proportion of prevalent topics is small in these major users,it helps to find the prevalent topics.更多还原