A Comparison of Implicit and Explicit Links for Web Page Classification一个比较隐式和显式链接的网页分类.pptVIP

下载本文档

7
0
约3.65千字
约 19页
2018-06-24 发布于四川
举报
版权申诉

A Comparison of Implicit and Explicit Links for Web Page Classification一个比较隐式和显式链接的网页分类.ppt

1、有哪些信誉好的足球投注网站（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

A Comparison of Implicit and Explicit Links for Web Page Classification一个比较隐式和显式链接的网页分类

A Comparison of Implicit and Explicit Links for Web Page Classification Dou Shen Jian-Tao Sun Qiang Yang Zheng Chen A presentation by Thomas Dinsdale-Young (some content from authors’ slides) Problem Classifying web pages Web is big (billions of pages) Make text mining easier Existing methods Manual (Web directories e.g. ODP, Yahoo!) Automatic Content-based: use the document Context-based: use hyperlink information Innovation: Implicit Links Web query logs contain user knowledge about relationships between pages and queries Entries: User Query Clicked results Time Leverage this knowledge for web page categorization Implicit Links Two types of implicit links: LI1: documents are linked if they were clicked by the same user for the same query LI2: documents are linked if they were clicked for the same query Explicit Links Between documents di and dj: LE1: There is a hyperlink to di from dj LE1: There is a hyperlink from di to dj LE1: There is a hyperlink between di and dj Link Statistics Classification Two approaches used: Classification by Linking Neighbours (CLN) Virtual Document-Based Classification Categories, training data and test data come from the Open Directory Project (ODP) Evaluating Classification Precision: Proportion of predicted class members that actually belong to the class Recall: Proportion of actual class members that are correctly predicted by the classifier F1-measure: Micro-average: each page equally weighted Macro-average: each category equally weighted Classification by Linking Neighbours Categorise a page with the most popular label among its linked pages Classification by Linking Neighbours Categorise a page with the most popular label among its linked pages Classification by Linking Neighbours Categorise a page with the most popular label among its linked pages Virtual Documents Add extra text to the document Local text: plain text and metadata Anchor text Extended anchor text Anchor sentence: sentences of a linked document containing que