Aprimorando o Desempenho de Sistemas de Recomendação Personalizados com Meta-Protótipos

Byron Leite Dantas BezerraFrancisco de Assis Tenório Carvalho

Recommender systems aim to furnish automatically personalized suggestions based on user preferences. These systems use information filtering (IF) techniques to recommend new items by comparing them with a user profile. This paper presents an approach where each user profile is modeled by a set of modal symbolic descriptions, which summarize the information given by the set of items already evaluated by the user. The comparison between a new item and a user profile is accomplished by a new suitable dissimilarity function which takes into account differences in content and position. This new approach is evaluated in comparison with the standard kNN method, which is an IF technique often used in this kind of system. Clique no link abaixo para buscar o texto completo deste trabalho na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: