摘要
[研究目的]用户生成文本包含丰富的需求信息与情感倾向,基于用户生成文本的关联规则识别,对量化用户情感需求、提升图书推荐的语义关联度具有重要的研究意义。[研究方法]通过融合读者标签的主题特征与情感评分,把模糊形式概念分析引入图书标签关联规则识别方法中,提出了一种基于用户生成文本的模糊关联规则识别方法(RFAR),该方法首先通过识别用户文本的主题特征,计算目标用户群的主题相似度,再利用文本情感分析确定用户与标签的情感依赖关系。最终,通过构建标签模糊概念格,实现模糊关联规则的在线识别。[研究结论]通过网络社区的用户数据验证了该方法的有效性,结果表明该方法能够实现用户需求的精准定位。
[Research purpose]User generated content(UGC)contains rich demand information and sentiment tendency.Association rule recognition based on UGC is of great significance for quantifying users'sentiment needs and improving the semantic correlation degree of book recommendations.[Research method]A novel book recommendation method based on fuzzy association rules was proposed by combining the topic features of readers'tags with sentiment ratings.The method first identifies the topic features of users'texts by calculating the topic similarity of target user groups.Afterwards,text sentiment analysis is used to determine the affective dependence between users and labels.Finally,the online identification of fuzzy association rules is realized by constructing a fuzzy concept lattice of labels.[Research conclusion]The validity of RFAR was identified through the user data of the network community,and the results showed that the method can locate the user needs accurately.
作者
张劲松
Zhang Jinsong(Library of Shandong Management University,Jinan 250357)
出处
《情报杂志》
CSSCI
北大核心
2021年第11期182-189,共8页
Journal of Intelligence
关键词
用户生成文本
图书标签
模糊关联规则
主题聚类
情感分析
相似度计算
user generated content
book labels
fuzzy association rules
topic clustering
sentiment analysis
similarity calculation
