基于用户信誉评级的UGC质量预判模型  

The Prediction Model of UGC Quality Based on User Credit Rating

作　　者：金燕[1] 闫婧[1] 

机构地区：[1]郑州大学信息管理学院,河南郑州450001

出　　处：《情报理论与实践》2016年第3期10-14,共5页Information Studies:Theory & Application

基　　金：国家社会科学基金项目"基于用户行为挖掘的UGC质量实时预判与控制机制研究"的成果之一;项目编号:15BTQ064

摘　　要：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 unbalanced quality of UGC, this paper proposes a prediction method of UGC quality based on user credit rating. By mining and analyzing the past behaviors of users for establishing, forwarding and discussing UGC in information ac- tivities, the paper builds a dynamic credit rating model of personal information behavior for users. According to the past credit rating of users, the paper predicts next UGC behavior of users as well as the UGC quality of the behavior. The experiment verifies that the method can present the quality of UGC behavior of different users in the observation period, in order to realize the real-time forecas- ting of UGC quality.

关 键 词：用户行为 用户生成内容 质量评价 模型 

分 类 号：G254[文化科学—图书馆学]