Research On Trajectory Data Mining Algorithm Based On LBSN

How to recommend for personalized users by using the obtained sparse data is a hot research topic in the area of big data and has wide application prospects.LBSN(Location-based social network)is a kind of social network.Different from the traditional social network,LBSN in addition to the traditional social network of human contact,but also can track and share the location information.For example,a mobile phone network is a typical LBSN.User can not only communication with other users,but also show his geographical location.In recent years,social networks have become increasely active,especially among young people.However,they are still based on the virtual world.However,with the development of wireless network and gps technology,it is easier to identify and share personal location infirmation.After adding a spatial dimention,social network in vitual world come back to the real word.Based on LBSN check-in data,This paper studies the following two problems in recommendation system:1)POI(Point of Interest)recommendation based on LBSN.2)Personalized traval route Recommendation based on LBSN.POI recommendation based on the LBSN provides users with personalized POI preference,such as attractions,hotels and shops and so on.A new POI recommendation model based on matrix factorization by considering the influences of both the geographical factor and the user factor,namely GeoUFM(Geographical and User Factorization Method),has been proposed in this paper.In GeoUFM,the objective function is defined by comparing the sorting results of both the objective function and the actual data.Experimental results on some real world LBSN data set have shown that GeoUFM obtained better performance on the recommendation precision and the recall rate.Personalized traval route Recommendation problem based on LBSN is to provide users with personalized POI sequence,in other word,it recommend user a route to traval.Based on the mining of personalized POI preference in GeoUFM,this paper proposes a personalized traval route recommendation model.In this framework of the model,the path construction is mapped into a tree construction process.The nodes in the tree are the geographical points of interest in the path,and the nodes of POI are got from the heuristic search in the candidated POI.The whole model takes into account the personalized preferences of user,the popularity of POI At the corresponding moment and time and space constraints,so that the recommendations close to the user’s preferences,and meet the constraints of the actual conditions.Through the test on the real data set,the experimental results show that the personalized path recommendation model has achieved good recommendation effect in the hit rate.