基于证据网络的在线商品评论有用性的评估模型  

An Evaluating Model for Online Product Review Helpfulness Based on Evidential Network

作　　者：王忠群 钱寅亮 叶安杰 陈云霞 Wang Zhongqun

机构地区：[1]安徽工程大学计算机与信息学院,安徽芜湖241000

出　　处：《情报理论与实践》2021年第1期154-161,共8页Information Studies:Theory & Application

基　　金：国家自然科学基金项目“C2C市场中基于行为树的销量识别与发布研究”(项目编号:71371012);教育部人文社会科学研究规划基金项目“基于知识图谱的网购商品评论可信性模型与评估方法研究”(项目编号:18YJA630114)的成果。

摘　　要：[目的/意义]评估并排序是缓解消费者难以发现有用性评论的有效途径。[方法/过程]基于评论有用性的影响因素间存在的多层次依赖关系,文章提出了一种基于证据网络的评论有用性的评估模型。实施以条件信度函数为参数的证据网络推理计算评论有用性,同时依据网络节点的计算值识别评论的有用性缺陷。[结果/结论]文章提出的模型不仅具有有效性,还可以识别评论的有用性缺陷,具备一定的评论有用性的可解释性。[Purpose/significance] Evaluating and ranking the helpfulness of online product reviews can make it easier for consumers to find out the useful online product reviews.[Method/process] Aiming at the multi-level dependency relations among influential factors of review helpfulness,this paper proposes an evaluating model for review helpfulness based on Evidential Network.Evidential Network reasoning using conditional belief function as parameters is carried out to compute the value of review helpfulness,and the defects in review helpfulness are identified according to the computing results of network node.[Result/conclusion] The empirical results show that our model is more effective,and can identify defects in review helpfulness,and has higher explanation of the review helpfulness.

关 键 词：在线商品评论 条件信度函数 证据网络 评论有用性 

分 类 号：F724.6[经济管理—产业经济] F274F426.63