| 摘要: |
| 锚文本对网络信息检索性能的提升作用已经得到验证,并被广泛地应用于商用网络搜索引擎.然而,锚文本制作的不可控性导致其中蕴含大量与目标网页不相关或具有作弊倾向的无用信息.另外,对于需要衡量检索结果服务质量的事务类查询,原始锚文本推荐的目标网页也往往与真实的用户体验不一致.为了解决上述问题,基于大规模真实用户的互联网浏览行为日志展开研究.首先提出了锚文本检索有效性的评估框架,然后分析了用户网络浏览点击行为与锚文本检索有效性之间的联系,挖掘了用户网络浏览点击行为中有助于筛选高质量锚文本的特征.基于这些特征,提出了两种超链接文档生成方法.实验结果表明,基于用户网络浏览点击行为特征筛选出的锚文本,与原始锚文本相比,能够明显地提升网络检索的性能. |
| 关键词: 用户网络浏览行为 锚文本 网络信息检索 |
| DOI：10.3724/SP.J.1001.2011.03873 |
| 分类号: |
| 基金项目:国家自然科学基金(60736044, 60903107); 高等学校博士学科点专项科研基金(20090002120005) |
|
| Retrieval Effectiveness Analysis for Anchor Texts |
|
ZHOU Bo1,2,3, LIU Yi-Qun1,2,3, ZHANG Min1,2,3, JIN Yi-Jiang1,2,3, MA Shao-Ping1,2,3
|
|
1.State Key Laboratory of Intelligent Technology and Systems, Tsinghua University, Beijing 100084, China;2.Tsinghua National Laboratory for Information Science and Technology, Tsinghua University, Beijing 100084, China;3.Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China
|
| Abstract: |
| Anchor texts have been extensively used and have proven to be effective over the years in commercial Web search engines; however, anchor texts are created freely without any supervision, which causes there to be a large number of noises and spam in anchor texts. Moreover, for transactional queries, which need measurements of service quality, destination, Web pages of anchor texts are not always consistent with user experience. To overcome these problems, this paper focuses on the large scale of the user Web browsing behavior data. This paper first proposes a framework for a retrieval effectiveness measurement. Then, based on this framework, analyze the relation between the user Web browsing behavior data and retrieval effectiveness of anchor texts is analyzed. The properties of a user Web browsing behavior, which are useful for Web search, are mined and quantified. Experimental results show that the proposed method offers an accurate evaluation on the retrieval effectiveness of anchor texts. |
| Key words: user Web browsing behavior anchor text Web information retrieval |
