| With the development of the information becoming electronic, website search become another increasingly important area after web search. Because ranking by content-based similarity faces many difficulties and link analysis technologies could not be directly applied to website search, there are many search work on website link analysis to enhance the query precision in website.Many of the research use the technology of ranking pages by content-based similarity, or some of them directly use link analysis technologies to the website search. But they all didn't work well. So some research work on the link analysis technologies for the website search.This paper first introduces the development and details of web link analysis technology, and then introduces some enhanced link analysis technology using in the website search. In order to enhancing the search prevision, this paper fully analysis the structure of the website and the characteristic of website search, import the conception of index page and article page and use them for the website search, then this paper present a website level link analysis algorithm basing on this conception. This algorithm uses the concetion of aothurity and hub in the HITS, ranks the pages in the website by aothurity value. In accordance with the website level link analysis algorithm, this paper present two methods to discriminate index pages and article pages, and it is important to the website information mining. In the experiment, this paper re-ranks the results of the full text search using the website level link analysis algorithm. It proved that the website level link analysis algorithm works very well for the website search compare to full text search and domain-restricted Google search. |
PDF Full Text Request