关键词: 主题爬行, 链接, TFIDF, 相似度

Abstract: In this paper, some experiments are conducted to analyze and verify the importance of table elements which lied in a Web page. In contr ast with Web pages, table elements can provide a large quantity of information ( beyond eighty percent) which is relevant to users’ information need. This conclusion can be utilizedto parse Web pages in the domain of focused crawling. After getting rid of elements except tables and headers, the efficiency of a focused crawler can be augmented distinctly and substantially.

Key words: focused crawling, URL, TFIDF, similarity