SpecificSearch : Un outil de recommandation automatique pour la veille d'information sur le web

The main goal of a content-based recommender system is to propose to a user new documents which are likely to have some interest for him considering feedbacks he gave for other documents. Different machine learning algorithms have been integrated to information retrieval systems in order to cope with this task. However, these systems have not become as popular on the web as the well-known search engines. Otherwise, the RSS feed aggregators allow to gather information on the web but these systems do not integrate machine learning in order to improve the quality of the recommendations with the users' feedbacks. Here, we present the functionnalities and the architecture of a web application available online called SpecificSearch (http://www.specific-search.com) which is an RSS aggregator integrating a machine learning algorithm. A first evaluation shows how a such tool can be implemented and how it can be useful. MOTS-CLÉS : recommandation automatique, veille d'information sur le web, filtrage adaptatif, flux RSS.