基于人类动力学的手机在线阅读用户的行为研究.docVIP

  基于人类动力学的手机在线阅读用户的行 为研究 王月，闫强** 5 10 15 20 25 30 35 40 （北京邮电大学经济管理学院） 摘要：基于手机在线阅读用户行为的数据统计，发现无论是从读者还是作品的角度分析，评 论时间间隔分布均可以用幂律函数很好的刻画。分布的幂指数会因用户的活跃度、作品的状 态等方面的不同而存在差异。用户行为的这种特征主要是由其兴趣变化引起的，用户的兴趣 会随时间推移慢慢衰减，但有可能因偶然事件的发生而突然回升，同时用户行为还会受作者 的行为、网站的经营模式等因素的影响。据此建立人类动力学模型，仿真重现用户行为的主 要特征，通过与实际数据的分析结果对比，验证了该模型能够比较准确的反应手机在线阅读 用户的行为特征。通过模型优化和参数调整，发现作者加快其更新作品的频率或者网站采取 个性化推荐的方式都可以增加用户的关注程度。 关键词：企业管理；人类动力学；手机阅读；幂律分布 中图分类号：F270 Studies on Users Behavior in Mobile Reading Online Based on Human Dynamics WANG Yue, YAN Qiang (Economic Management School, Beijing University of Posts and Telecommunications, Beijing 100876) Abstract: Based on the data statistics of mobile reading users’ behavior, this paper found that no matter analyzing from the perspective of the readers or the books, the distribution of time interval between two consecutive comments is a power-law with heavy-tailed. The power exponent is different because of the difference of users’ active degree and work state, etc. The users’ behavioral characteristic is caused by the change of their interest. The interest will reduce as time goes on and rise again because of some accidental event. The behavior is also influence by the writers and the website. Based on these, this paper built a human dynamics model impacted by the interest. The simulation result agrees with the empirical result well, which prove that this model could reflect the mobile reading users’ behavioral characteristic. Through the parameter adjustment and model optimization, it is found that if the authors can speed up the rate of update their works or the website can take the approach of personalized recommendation, more readers could been attracted. Key words: business management; human dynamics; mobile reading; power-law distribution 0 引言 随着终端设备的更新和网络技术的进步，移动互联网上的业务和应用也在不断的发展， 手机阅读就是其中发展较快的一种。易观国际 Enfodesk 产业数据库发布的报告显示，2010 年第 4 季度手机阅读活跃用户数已达 23290 万人，市场收入达 94895 万元，已经成为移动互 联网中的重要应用。手机在线阅读是手机阅读的一种重要形式，我们可以认为手机在线阅读 是以手机为载体的网络文学，因此它融合了网络文学和手机阅读的特点，包括公开性