摘要
[目的/意义]从社会化标签视角探讨图书推荐系统模型。[过程/方法]分析3种主流的标签推荐模型(基于主题、网络和张量的模型)和3种标签推荐算法(基于内容、协同过滤和混合推荐算法)的优缺点,构建基于标签的图书推荐系统模型。[结果/结论]该模型由数据采集模块、数据预处理模块和协同过滤推荐模块组成,可以在一定程度上解决推荐系统中的标签语义模糊、用户兴趣多变、可扩展性差等问题,以提高图书推荐效率和准确度。
[Purpose/significance]The paper is to discuss book recommendation system model from the perspective of social tags. [Method/process]The paper analyzes the advantages and disadvantages of three mainstream recommendation models （subject, network, and tensor based models） and three tag recommendation algorithms （content, collaborative filtering based and hybrid algorithms）, and constructs a tag-based book recommendation system model. [Result/conclusion]The model is composed of data acquisition module, data preprocessing module and collaborative filtering recommendation module, can solve the problems of fuzzy tag semantics, changing users＇ interest, and poor expandability to some extent, so as to improve efficiency and accuracy of book recommendation.
出处
《情报探索》
2016年第10期80-85,共6页
Information Research
基金
天津师范大学教育基金项目"以科研为导向的图书馆学科服务平台建设"(项目编号:52WT1408)研究成果之一
关键词
社会化标签
图书推荐系统
LDA
协同过滤
social tag
book recommendation system
LDA
collaborative filtering
