IRIT at TREC 2002: Web Track

In Trec-2002, we participated in the Web Trec (named page finding task). There a re two k inds of information that can b e used while finding the e xpected p age, content i nformation and link information. We exploited both of them. That is to say, our system i s content-based and link-based. As to link information, we only used anchor text and connections, and topology between pages is ignored. We submitted two runs. One is based on traditional contented-based retrieval, the other try to combine content-based retrieval and link-based retrieval t o g et better result.