计算机科学与探索 ›› 2012, Vol. 6 ›› Issue (8): 717-725.DOI: 10.3778/j.issn.1673-9418.2012.08.005

• 学术研究 • 上一篇    下一篇

在线百科间的标签推荐算法

刘  阔1,姚舒扬2,邓志鸿2,3+   

  1. 1. 北京大学 信息科学技术学院 计算机系,北京 100871
    2. 北京大学 信息科学技术学院 智能科学系,北京 100871
    3. 北京大学 信息科学技术学院 机器感知与智能教育部重点实验室,北京 100871
  • 出版日期:2012-08-01 发布日期:2012-08-06

Tag Recommendation among Different Online Encyclopedia Systems

LIU Kuo1, YAO Shuyang2, DENG Zhihong2,3+   

  1. 1. Department of Computer Science, School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China
    2. Department of Machine Intelligence, School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China
    3. Key Laboratory of Machine Perception (Ministry of Education), School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China
  • Online:2012-08-01 Published:2012-08-06

摘要: 信息社会中在线百科已成为人们获取知识的重要途径,而在线百科的标签系统作为其重要组成部分,不仅可以帮助人们在浏览某张页面时获取其他相关页面的信息,而且对于海量文本分类,以及提高在线百科检索系统的检索效率都有很大帮助。充分利用在线百科页面间的链接关系,提出了一种基于页面间的同质性原理和向量空间模型的全新针对在线百科的标签推荐算法HVSM(homogeneous principle based vector space model)。该标签推荐算法具有普适性,可在不同在线百科系统间推荐标签。实验结果表明,通过与朴素推荐算法NAM(na?ve recommendation model)进行比较,新的推荐算法可以达到更高的准确率。并且通过对实验数据进行分析,得到了若干有益的结论,为今后的研究工作奠定了基础。

关键词: 在线百科系统, 标签推荐, 同质性原理, 向量空间模型

Abstract: Online encyclopedia systems have become an important way for people to acquire knowledge. As an important part of online encyclopedia systems, tags of each page not only help users get more relevant information for further reading, but also enhance the efficiency of the retrieval system. After taking full advantage of linkage relations between pages of online encyclopedia systems, this paper proposes a new method, HVSM (homogeneous principle based vector space model), which is universal and aims at recommending tags between different online encyclopedia systems. Experimental results show that this method can achieve a good accuracy which is higher than NAM (na?ve recommendation model). Some useful conclusions are obtained through the analysis of experimental results, which lay a solid foundation for further research.

Key words: online encyclopedia system, tag recommendation, homogeneous principle, vector space model