您的位置:首页  > 论文页面

基于语义概念相似度的科技文献推荐算法

发表时间:2009-02-28  浏览量:2046  下载量:965
全部作者: 王位春,张铭
作者单位: 北京大学信息科学技术学院
摘 要: 提出了一种基于语义概念相似度的科技文献个性化推荐方法。首先使用用户对各文献标记的标签(tag)构建语义概念,然后使用构建的语义概念表示用户偏好和文献特征(profile),并根据profile的相似度选择相邻用户,最终在相邻用户标记过的文献中通过基于文本过滤的推荐算法选择相关文献推荐给用户。经实验验证,该方法运用于所在科研服务平台的个性化推荐模块中,不仅提高了推荐的准确率,而且能为用户发现新的感兴趣的资源。
关 键 词: 计算机应用;个性化推荐;协作过滤;语义概念;标签
Title: A recommendation algorithm for science and technology literature based on semantic concept similarity
Author: WANG Weichun, ZHANG Ming
Organization: School of Electronics Engineering and Computer Science, Peking University
Abstract: This paper presents a personalized recommendation algorithm for science and technology literature based on semantic concept similarity. It first uses user tags to build semantic concepts, and uses these concepts to represent the profiles of users and items. Based on the similarity of profiles, neighbor users are selected and content-based text filtering is used to extract recommendation items from these neighbor users’ tagged items. This approach is implemented in the personalized recommendation module of scientific literature service platform. Through experimental verification, the approach not only improves the accuracy of recommendation, but also helps the users find the new interesting items.
Key words: computer application; personalized recommendation; collaborative filtering; semantic concept; tag
发表期数: 2009年2月第4期
引用格式: 王位春,张铭. 基于语义概念相似度的科技文献推荐算法[J]. 中国科技论文在线精品论文,2009,2(4):372-378.
 
1 评论数 0
暂无评论
友情链接