| With the development of the Internet,information overload problem make it diffcult to get satisfactory resource from a great deal of Internet information.The emergence of the recommendation system can effectively help people select information that they are interested in.It is a pity that traditional recommender system rarely use the location information about users in system.Moreover,conventional system hardly make personalized recommendation.The development of mobile devices and wireless communication make it easy to acquire users' locations.Thus,location-aware recommendation has attracted more and more attention.Because of the additive about the location information of users,location-aware recommendation system can mini a user's preference more accurately.So far,location-aware system is divided in two types according to the location information.In the actual scene,the data of record about users in system is sparse because a user usually has checked at a few locations.This dissertation studies the tradition recommender system and location-aware recommendation system,describing the common recommendation algorithms,the methods of location-aware and fusion strategies.Moreover,we propose two ways based on multiobjective optimization to realize the recommender system.One is location-aware media recommendation based on multi-objective optimization,the other is location recommendation based on multi-objective optimization.The first one improve the accuracy and coverage of recommender system by multi-objective optimization.The latter one model the common interest and the individual interest of a user using multi-objective optimization.In this dissertation,we compared the improved algorithm with other recommendation methods by some experiments and verify the feasibility and the effectiveness of our algorithm.The results of experiments show that our proposed methods can recommend satisfactory items and has a good performance of alleviating problems. |
PDF Full Text Request