Page Ranking 2007 04 18 Page Rank Sergey

检索排序算法 Page Ranking 2007 -04 -18

Page Rank 历史 • Sergey Brin和Lawrence Page在 1998年提出了Page. Rank 算法，同年J. Kleinberg提出了HITS算法 • Lawrence Page, Sergey Brin, Rajeev Motwani, Terry Winograd, 'The Page. Rank Citation Ranking: Bringing Order to the Web', 1998, http: //www-db. stanford. edu/~backrub/pageranksub. ps • 为了更高效地计算 Page. Rank, 以下是改良以后的一篇论文。 Taher H. Haveliwala, ‘Efficient Computation of Page. Rank’, Stanford Technical Report, 1999, http: //dbpubs. stanford. edu: 8090/pub/1999 -31 • Page. Rank(TM) 是美国 Google 公司的登记注册商标。

Computing Page. Rank - initialize vector over web pages loop: - new ranks sum of normalized backlink ranks - compute normalizing factor - add escape term - control parameter while - stop when converged

Page. Rank实例 链接源I D 1 2 3 4 5 6 7 链接目标 ID 2, 3 , 4, 5, 7 1 1, 2 2, 3, 5 1, 3, 4, 6 1, 5 5

Page. Rank实例 A=[ ] 1, 2, 3, 4, 5, 6, 7 0, 1, 1, 0, 1; 1, 0, 0, 0; 1, 1, 0, 0, 0; 0, 1, 1, 0, 0; 1, 0, 1, 0; 1, 0, 0, 0, 1, 0, 0;

Page. Rank实例 • 流入量 ＝(ID=2发出的Rank) +(ID=3发出的Rank) +(ID=5发出的Rank) +(ID=6发出的Rank) = 0. 166+0. 141/2+0. 179/4+0. 045/2 = 0. 30375

ODP • The Open Directory Project is the largest, most comprehensive human-edited directory of the Web. It is constructed and maintained by a vast, global community of volunteer editors. • The Open Directory provides the means for the Internet to organize itself. As the Internet grows, so do the number of net-citizens. These citizens can each organize a small portion of the web and present it back to the rest of the population, culling out (挑选出）the bad and useless and keeping only the best content.

The Open Directory was founded in the spirit of the Open Source movement, and is the only major directory that is 100% free. There is not, nor will there ever be, a cost to submit a site to the directory, and/or to use the directory's data. The Open Directory data is made available for free to anyone who agrees to comply with the free use license.

The frame of system using topic-sensitive Page. Rank

Learning to Rank • In most recent works, learning to rank is defined as having the following two properties: • Feature based; • Discriminative (有识别能力）training. • Automatic parameter tuning for existing ranking models.