Melhorias para um sistema de recomendação baseado em conhecimento a partir da representação semântica de conteúdos

Abstract

The Recommendation systems are already established as tools that support users to overcome the difficulties caused by the excessive volume of content available in digital format and was designed to conduct automated the content classification tasks and relationship of this with wins users. One of the problems observed in these systems is related to the weakness of some classification approaches and content relationship rely mainly on methodical aspects of the discussed subjects. Recommendation systems based on knowledge seek to mitigate this problem from the incorporation of semantic elements in the indexing processes and material relationship. Despite good results observed, research needs are also identified, both used to classify content activities, such as the representation and treatment of relationships between content and potential stakeholders. This paper seeks to contribute to the development in this area from the proposal for a recommendation system based on knowledge and facing the recommendation of educational materials in a context of small groups of students. The spread of this system is through a semantics of the merger process associated with these types of concerns and also with the use of semantic aspects to represent the needs and relationships originated by system users. The main distinguishing feature of this system is located in the use of a hybrid recommendation algorithm in which both syntactic and semantic aspects are employed. To evaluate the proposed recommendation system, it is due for prototyping and testing in a controlled environment.