An adaptive correlation-based group recommendation system

An adaptive correlation-based group recommendation system or recommender is proposed which takes the members´ interactions into account. Thus, the proposed method can predict members´ relationships by acquiring the correlations between a group and its members. With the predicted correlations, each member´s weight can be estimated. Finally, the group´s ratings of those non-rated items can be predicted by merging members´ ratings and their estimated weights. Without complex computation, the proposed group recommendation system can achieve accurate predictions of the group´s ratings. Therefore, it suits for helping a group to make decisions since its members have different opinions for selecting multimedia contents.