Mechanism
As a social media platform, Weibo’s Real-time Trending List has got certain social impacts undoubtedly. Our socio-technical diagram analyzes the relations among Weibo users, Sina Company Team and Weibo’s Technical System.
On the front-end, users including company official accounts, governmental official accounts, personal accounts, and paid/voluntary supporters keep attention to the Weibo Real-time Trending List—they search for the trending, as well as repost and comment on the contents, which will lead to hot discussion. This is how Weibo Real-time Trending List has influenced on the users, but in return, users’ opinions and behavior provides it with feedback. For example, user searching for certain trending and comments on it, and other users will look through, from which the interaction arisen around the trending topics happens. At the same time, users’ behavior like searching for different topics is reflected as changes on Weibo Real-time Trending List. On the other hand, all users will give feedback to the Sina Company Team if they have problems with the Weibo Real-time Trending List, such as user interface, which will be fixed by its Operating Team and Technical Team.
On the back-end, all the data provided by Weibo users fill in the data-pool as candidate topics for the list. After being inputted into the filter, those content will be filtered by certain scheme searching for sensitive words like exotic terms carried out by machine automatically and manual work. Then the residuals will be inputted into the Ranking Algorithm that will be explained in the architecture part in order to generate Weibo Real-time Trending List. The whole Technical System is developed by the Technical Team of Sina Company Team.
To be specific, the architecture explains the technical mechanism of how Weibo Real-time Trending List is shaped and maintained. Firstly, users’ behavior in the front-end will be imputed into a counting loop that counts for the number of reposting, liking, searching, and commenting under each topic separately without weight, after which it forms a list of topics without order. And then, the data will be inputted into Bayesian Average formula: WR=[V/(v+m)]R+[m/(v+m)]C, in which the number got from last steps will be computed with different weight set by Weibo Company and presented as a pre-ranking list in descending sort. Then this list will be adjusted based on post time, which follows the concept of Newton’s Law of Cooling: , so that the newly coming topics can have the chance appearing on the List. So far, Weibo Real-time Trending List has basically got its form, but that is not the end. As the Weibo Real-time Trending List is viewed, reposted, liked and searched by the users constantly, this loop will keep running again and again.
