Research On Geo-Social And Textual K Nearest Neighbor Queries In Road Networks

With the development of social networks,the volumes of available social network data are growing rapidly.For example,Facebook had on average 654 million active mobile users on a daily basis before June 2014.This type of data is mostly a product of human communication,and it offers tremendous opportunities for predictions and recommendation,since users with social links influence each other and may have similar preferences.For instance,a group of Greek Americans who are friends may all be interested in a Greek restaurant.Furthermore,the web objects that represent services,e.g.,stores,hotels,and tourist attractions,are increasingly being geo-tagged.In this paper,we provide support for two new types of query,the Geo-Social Textual k Nearest Neighbors(GSTkNN)query in Road Networks and the Maximizing Bichromatic Reverse Geo-Social Textual k Nearest Neighbors(MaxBRGSTkNN)query in Road Networks.The GSTkNN query takes into account spatial,textual,and social information,and recommends geotagged objects to users.The MaxBRGSTkNN query takes into account spatial,textual,and social information,and finds the location and the text contents to include in an advertisement so that it will be displayed to the maximum number of users.To address this,we propose a hybrid index,the GIM-tree,which indexes locations,texts,and social information of geo-tagged users and objects,and then,using the GIM-tree,we present efficient GSTkNN query and MaxBRGSTkNN query processing algorithms that exploit several pruning strategies.The effectiveness of GSTkNN and MaxBRGSTkNN retrieval is characterized via extensive experiments using real datasets offer insight into the efficiency of the proposed index and algorithms.