手机在线阅读中人类行为动力学的研究

发布时间：2018-06-05 13:05

本文选题：人类动力学 + 幂律分布　；参考：《北京邮电大学》2012年硕士论文

【摘要】：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。虽然手机阅读的用户基础好、市场潜力大,但目前仍旧缺乏清晰的盈利模式,用户付费意愿不强,造成手机阅读业务收入不高的现状。本文用人类动力学的理论和方法研究手机在线用户的行为特征,分析影响其阅读行为的主要因素,并且针对性的提出利于手机在线阅读发展的合理化建议。基于手机在线阅读用户行为的数据统计,发现用户的行为除了具有周期性、相关性等基本特征之外,其胖尾特征也十分明显,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,用户活跃度越高,其幂指数越大,相对于完结作品,连载作品更能够引发频繁的评论,而用户不会因为作品收费而减少对它的评论或点击,说明对于用户来说,能吸引他们的往往是质量高、更新快的作品,“免费的午餐”不能赚取关注度。用户行为的这些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,因此建议网站优化推荐方式,并邀约高水平的写手入驻。
[Abstract]:With the update of the terminal equipment and the progress of the network technology, the business and application on the mobile Internet are developing constantly, and the mobile phone reading is one of them. Although the mobile phone reading user base is good, the market potential is big, but at present still lacks the clear profit pattern, the user pays the will is not strong, causes the handset reading service income not to be high the present situation. In this paper, we use the theory and method of human dynamics to study the behavior characteristics of mobile phone online users, analyze the main factors that affect their reading behavior, and put forward some reasonable suggestions for the development of mobile phone online reading. Based on the statistics of mobile phone online reading user behavior, it is found that in addition to the periodicity, correlation and other basic characteristics of user behavior, its fat-tailed features are also very obvious, whether from the perspective of readers or works. The distribution of comment time interval can be well characterized by power law function. The power exponent of the distribution will vary according to the user's activity, the state of the work, etc. The higher the user's activity, the greater the power exponent. Compared with the finished work, serial works are more likely to trigger frequent comments. The fact that users don't comment or click on a work for a fee means that it's often high quality, fast-updated works that appeal to users, and "free lunches" don't earn a lot of attention. These characteristics of user behavior are mainly caused by changes in their interests, which slowly attenuate over time, but may suddenly pick up as a result of accidental events, while user behavior is also subject to the behavior of the author. Website business model and other factors. Based on this, the human dynamics model is established, and the main characteristics of user behavior are simulated and reproduced. By comparing with the analysis results of actual data, it is verified that the model can accurately reflect the behavior characteristics of mobile phone online reading users. Through model optimization and parameter adjustment, it is found that the author can increase the attention of the user by speeding up the frequency of updating his work or the website adopting personalized recommendation, so it is suggested that the website should optimize the recommendation mode. And invite a high level of writers to enter.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：C912.6

【引证文献】