Use noisy link analysis to improve web search

ABSTRACT

Link-based ranking algorithm is very important for current success and popular of Web Search Engine. In order to get high rank, some try to improve contents of web pages while others just put dirty tricks, such as link spam. Link spam is a trick targeting at link-based ranking algorithms by artificially created tight link structures to push some target pages get undeserved high ranks. This problem becomes even worse with the advent of wikis, blogs, forums, which are rich in links. We tackle the problem of improving link-based ranking by more fundamental viewpoint--"noisy link" analysis. Motivated by how "non-voting" hyperlinks affect quality of ranking, we propose an approach and corresponding penalty strategies to both detect and handle "noisy link" effectively and automatically. We also compared our approach with other related works to demonstrate that our approach is rather effective in noisy link filtering and could improve the final search results significantly.

REFERENCES

• 1: André Luiz da Costa Carvalho , Paul - Alexandru Chirita , Edleno Silva de Moura , Pável Calado , Wolfgang Nejdl, Site level noise removal for search engines, Proceedings of the 15th international conference on World Wide Web, May 23-26, 2006, Edinburgh, Scotland [doi>10.1145/1135777.1135793]
• 2: Xiaoguang Qi , Lan Nie , Brian D. Davison, Measuring similarity to detect qualified links, Proceedings of the 3rd international workshop on Adversarial information retrieval on the web, May 08-08, 2007, Banff, Alberta, Canada [doi>10.1145/1244408.1244418]
• 3: Zoltán Gyöngyi , Hector Garcia-Molina , Jan Pedersen, Combating web spam with trustrank, Proceedings of the Thirtieth international conference on Very large data bases, p.576-587, August 31-September 03, 2004, Toronto, Canada

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval
          Subjects: Information filtering

General Terms:
Algorithms, Experimentation

Keywords:
link penalty, link-based ranking, noisy link, pagerank