This thesis addresses the problem of analysis the Web hyperlink structure in order to locate authoritative web pages relevant to a certain query subject. In particular, we discuss the HITS algorithm, a pioneering method that exploits the web link structure in order to locate important information sources, as well as other algorithms based on HITS such as ARC and Page Rank. We experiment with different aspects of HITS, and introduce new heuristics for assigning the link weights based on textual contexts.