Spatial-textual Skyline Query

With the development of mobile network,location-based services are becoming more and more popular,and a large number of interest points with textual information and spatial information are gradually increasing.In order to return satisfactory results from plenty of interest points,the user's query usually have a certain preference.Skyline query as a very important preference query,scholars have conducted extensive research.Considering the user's direction and social information generated by social network,in this paper,we study the spatial-textual skyline query further.Firstly,according to the user's direction,this paper propose a direction-based spatial-textual skyline query algorithm.In this algorithm,direction is applied to the spatial-textual skyline query and skyline objects located at different directions are returned for the user.The determination of skyline object depends on three aspects: location proximity,textual similarity and direction.Two functions are introduced to compute the textual relevance and spatial proximity respectively.In order to improve the query speed,effective pruning strategy and termination theorem are proposed,.Secondly,on the basis of the user's choice often affected by friends,using the social information generated in social network,a new query algorithm is proposed,which is called social-based spatial-textual skyline query.Identifying the skyline object depends on three aspects:spatial distance between the object and the query user,textual similarity of the keywords and social relevance.One function is introduced to compute the social relevance.In order to improve the satisfaction of the query,we extend the query,and propose a constrained social-based spatial-textual skyline query.In addition,a novel concept of constrained skyline is introduced.For each query,the application of pruning strategies and termination conditions are put forward to improve the query efficiency.Finally,empirical results on data set show that our proposed methods are effective and efficient.