论文部分内容阅读
针对目前UGC质量不均衡的问题,提出一种基于用户信誉评级的UGC质量预判方法。通过挖掘、分析用户以往信息活动中的UGC创建、转发、评论等历史行为,为用户建立起个人信息行为动态信誉评级模型。根据用户过往的信誉等级,预判用户下一次UGC行为及该行为所产生的UGC的质量。实验证明该方法能够呈现不同用户在观察期的UGC行为质量,实现对UGC质量的实时预判。
Aiming at the problem of unbalanced quality of UGC, a UGC quality prediction method based on user reputation rating is proposed. Through excavating and analyzing historical behaviors such as creating, forwarding and commenting UGC in the past information activities of users, a dynamic credit rating model of personal information behavior is established for users. According to the user’s past credit rating, predict the next UGC behavior of the user and the quality of the UGC generated by the behavior. Experimental results show that this method can show the UGC behavior quality of different users during the observation period and realize the real-time prediction of UGC quality.