基于用户交互模型的会员网站每日活跃用户量分析

现有的网站往往向注册用户提供服务，而网站中的每日活跃用户量往往决定着网站建设的成败。为了描述网站中每日活跃用户的动态变化，本文将用户在网站上的交互行为分为响应、扩散和衰落三种类型，并提出了一种基于用户交互模型的每日活跃用户量模型。通过响应、扩散和衰落三种交互行为的定义，进一步推导出每日活跃用户随着时间的变化率。通过对每日活跃用户随着时间的变化率进行分析得到如下结论，当网站中用户的响应概率小于衰落概率时网站的活跃用户将趋向于0，反之活跃用户趋于某个固定的常数。大量的真实数据实验表明，不论网站自身的运营是否成功，本文提出的方法都可以很好的描述网站的每日活跃用户数量及其发展趋势，这对于网站的建设和维护具有重要的指导意义。

Analysis of Daily Active User in Membership-based Websites Based on User Interaction Model

Current websites usually provide services only for registered users, and the number of daily active user in each website decides whether the website will succeed. In order to describe the dynamics of daily active users in websites, this paper divided actions between users in a website into reaction, diffusion and decay, and proposed a user interaction based daily active user model. According to the definitions of reaction, diffusion and decay, we further inferred the rate of change of daily active users to time. While analyzing the rate of change of daily active users to time, we concluded that the number of daily active users would asymptotically tend to be a constant, and even zero while the rate of reaction was less than the rate of decay. Massive real dataset experiments show that, no matter whether the website itself is succeeded or not, the proposed approach can effectively describe the number of daily active users and its trend, and then can provide guidelines for building and maintaining of websites.