欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
根据评论内容的特性,提出了一种基于主题-对立情感依赖模型（topic-opposite sentiment dependency model,TOSDM）的虚假评论检测方法。首先构建TOSDM模型,利用该模型提取评论的主题信息以及主题对应的情感信息;然后结合评论的主题以及情感信息,分析并提取6维评论内容特征;最后利用这些评论内容特征,采用有监督学习的分类器对虚假评论进行检测。在大众点评网获取的2009-2012年的5个领域的评论数据集上进行了实验,实验表明,提取评论主题信息以及主题对立情感信息能够提高虚假评论的检测效果,TOSDM的虚假评论检测效果优于其他模型。
英文摘要：
Based on the content characteristics of the review,we proposed a method to detect fake reviews based on topic-opposite sentiment dependency model（TOSDM）.We construct TOSDM of reviews based on sentiment dependency,which is used to extract review topic information and their corresponding emotional information.Moreover,We analyze and extract 6types of review content features combined review topics and emotional information,and exploit all these review content features which were extracted in the back step to detect fake reviews using classifier based on supervised method.The experiments are completed with the reviews of 5fields from Dianping.com in 2009-2012.The experimental results show that extracting the reviews topic information and topic-opposite sentiment information can improve the detection effect for fake reviews,and TOSDM outperforms other generative models.
同期刊论文项目
同项目期刊论文
Restricted domain question-answering text retrieval method based on supervised latent dirichlet allo
Approaches to Detect Micro-blog User Interest Communities through theIntegration of Explicit User Re
期刊信息
位置：