Can't get more satisfaction?: game-theoretic group-recommendation of educational resources

Students' satisfaction from educational resources is a subjective perception of how well these resources meet students' expectations for learning. Recommending educational resources to groups of students, targeting at optimizing all students' satisfaction, is a complicated task due to the lack of joint group profiles. Instead of merging individual profiles or fusing individual recommendations, this paper follows a game-theoretic perspective for solving conflict of interest among students and recommending resources to groups in online collaborative learning contexts: the group members are the players, the resources comprise the set of possible actions, and maximizing each individual member's satisfaction from the selected resources is a problem of finding the Nash Equilibrium. In case the Nash Equilibrium is Pareto efficient, none of the players can get more payoff (satisfaction) without decreasing the payoff of any other player, indicating an optimal benefit for the group as a whole. The comparative evaluation of the suggested approach to other state-of-the-art methods provided statistically significant results regarding the error in predicted group satisfaction from the recommendation and the goodness of the ranked list of recommendations.