Abstract
There is no doubt that the World Wide Web has made easier the task of searching for information on the Internet. The amount of information obtained (some of them irrelevant ones) increases day after day and creates opportunities for a new breed of systems named "Recommender Systems". These systems have emerged as one successful approach to tackle the problem of information overload. Traditional recommender systems suggest research items using well-known text mining techniques, however they fail when there are no identical keywords to match searches. In order to overcome this and other limitations, several studies have been made in order to verify the benefits of ontology-based approaches to create what is known as ontology-based recommender systems. This paper analyzes several ontology-based recommender systems and discusses some classification criteria in order to define a common architecture for these special types of recommender systems. The architecture presented is discussed in details and recommended as basis architecture for other ontology-based recommender systems.