Multi-Objective Optimization Algorithms For Point-of-interest Route Recommendation In LBSN

How to recommend for personalized users by using the obtained information is a hot research topic in the information era,which has wide application prospects and extensive research value.LBSN is the Location-based social network.Different from the traditional social network,LBSN can not only connect people in traditional social networks,but also track and share people's location information.For example,mobile communication equipment can be seen as LBSN.Users can not only communicate with other users by using them,but also can share their location.Nowadays,social networks have become increasely popular among young people,which are still based on the online network at early stage.However,with the developement of wireless communication and GPS technology,it is easier to identify and share their location information.After adding a geospatial dimension,social network in online virtual world are connected to the real word.Based on real dataset,this thesis studies the following problems in recommendation system:1)Region recommendation based on GPS dataset.2)Personalized POI route Recommendation based on LBSN dataset.Hot region recommendation model based on the GPS trajectory data provides users with personalized region recommendation,such as hotel,park,shopping center and so on.A new region recommendation model based on matrix factorization has been proposed in this paper.In this model,first,it analyzes the different features of the region that users may be interested in and then uses a density-based hot-region discovery method to search for hot regions that different users may frequently visit.After that,by calculating the interest degree of the currently known region,a matrix factorization method is used to predict the interest degree of the regions that users have not been visited.In this way,regions with high interest degree are selected to recommend to users.Experimental results on some real-world GPS dataset have shown this model obtained better performance on the recommendation precision and MAE value.Personalized POI route recommendation problem based on LBSN is to provide users with some personalized POI sequences.In other word,it recommends user a POI route.A personalized POI route recommendation model based on multi-objective optimization model is proposed by mining the users' preference based on real LBSN check-in dataset.In this POI route recommendation framework,a multi-objective optimization model is adopted to comprehensively consider the user's personalized interest preferences,location popularity,distance and time constraints.And then several corresponding personalized POI routes are constructed through multi-objective optimization algorithms.The recommendation results are more in line with the user's preferences and meet the constraints.Through the test on the real dataset,the experimental results show that the personalized POI route recommendation model has achieved better recommendation performance in the hit rate.