Sistema de Recomendación para la Recuperación Automática de Enlaces Web Rotos

In theWebpages accessed when navigating throughInternetor even in our ownWebpages, we sometimes find links which are not valid any more. The search of the rightWebpages which correspond to those links is often hard. In this work we have analyzed di!erent sources of information to automatically recover brokenWeb links so that the user can be o!ered a list of possible pages to substitute that link. Specifically, we have used either the anchor text or theWebpage containing the link, or a combination of both. The information extracted is then used to perform a search with some of the usual search engines, such asGoogleorYahoo. The candidate pages are then ranked applying information retrieval techniques on their content. Finally, the user is presented the pages resulting from this process. We report the analysis of a number of issues on a set of links randomly chosen, what has allowed us to decide the conditions under which the system can make a reliable recommendation.