Recommender Systems have been a key issue in the development and success of electronic commerce. These websites usually provide hundreds or thousands of products related with a simple user query. Although the wide range of products obtained as a responde can initially be considered an advantage, it sometimes becomes rather a drawback, since customers have to find, among a high number of products, those ones they are really interested in. In many cases, customers can not explore so many alternatives, and they finally choose a product quickly, which meets his needs moderately, or they even give up with their search. In order to solve these problems several techniques have been proposed, such as Recommender Systems. The aim of this kind of software is to support users in their search processes to lead them towards interesting products by means of recommendations, showing them a list of products ordered according the degree to which they meet their needs. All the recommendation systems have the same aim: guiding users by means of recommendations towards those products which are the most adequate for them. However, recommender systems may use different techniques or algorithms in order to generate the recommendations. According to these techniques we can classify recommender systems as: demographic recommender systems, content-based recommender systems, collaborative recommender Systems, knowledge-based recommender Systems, hybrid recommender systems, etc.

Although at SINBAD² we have carried out studies and implementations about different types of Recommendation Systems, our major effort has been focused in the Recommender Systems based on Knowledge and processes to rebuild good user’s profiles based on small amounts of information, in order to get better recommendations.