基于群体智能挖掘的个性化商品评论呈现方法

CrowdReview:personalized product review presentation based on crowd intelligence mining

为了研究基于群体智能挖掘的个性化商品评论呈现方法,以大众点评用户评论为研究对象,对大众点评中的用户评论进行特征提取,并发现兴趣相似的用户.特别是提出基于情感特征和主题分布特征的相似群体发现方法,通过提取用户历史评论的情感特征和主题分布特征,刻画用户之间情感和主题的相似度,并发现兴趣相似的用户群体,实现个性化评论呈现.实验结果表明,采用提出的方法可以体现用户间兴趣的相似性并发现与用户有相似兴趣的群体,向用户个性化呈现评论.

Personalized product reviews were presented by exploring the public comment data in order to study the method of personalized product review presentation based on crowd intelligence mining.Sentiment feature and topic distribution feature from user reviews were extracted and users were clustered into different groups based on the sentiment similarity and topic distribution similarity of their reviews.Experimental results show that our approach can reflect the similarity of users and find the same group.For a given user,reviews can be only presented from those who belong to the same group as oneself.