In the web pages accessed when navigating throughout Internet, or even in our own web pages, we sometimes
find links which are not valid any more. The search of the right web pages which correspond to those links
is often hard. In this work we have analyzed different sources of information to automatically recover broken
web links so that the user can be offered a list of possible pages to substitute that link. Specifically, we have
used either the anchor text or the web page containing the link, or a combination of both. We report the analysis
of a number of issues arising when trying to recover a set of links randomly chosen. This analysis has allowed us to
decide the cases in which the system can perform the retrieval of some pages to substitute the broken link. Results
have shown that the system is able to do reliable recommendations in many cases, specially under certain conditions on
the anchor text and the parent page.