Evaluating the importance of Web pages

With die enormous growth of Web pages, an increasing number of users are relying on Web search engine. But now it is impossible for spider to retrieval all pages on the Web. Furthermore, because of replicated contents the Spider needn't to retrieve all Web pages. Apparently, visiting important pages firstly can be an efficient method. So evaluating the authority of Web pages has becoming a key problem. Moreover, if we can compute die authority of all pages appropriately their the search engine can return important pages when processing the queries. Then users can obtain satisfying results. Google is one of the best search engines. It can produce much more satisfying search results than many other existing systems. It has a great progress on both precision and recall based on hyperlink analysis. This paper analyzed Google's Algorithm on PageRank in details and presented some disadvantages of this algorithm, for instance, preferring old pages, ignoring special sites and inaccurate judge of hyperlinks pointed out from one page. Furthermore, we describe our improved algorithm. Experiments show that our consideration on evaluating the importance of pages can make an improvement over the original algorithm.