A Survey on Nearest Neighbor Search with Keywords

Abstract

Today many applications use a new forms of query called as spatial keyword query which include finding objects closest to a specified location that contains specific set of keywords. For example, the nearest hotels to a specific location that contain facilities lunch and dry cleaning. Such query would ask for the hotels that are closest among those which provides facilities free lunch and dry all at the same time instead of considering all the hotels. Currently using IR2-tree is the best solution to such queries, which has a few deficiencies that seriously impact its efficiency. In this paper, we present a review on various methods used for NN search with keywords. whether a point is in rectangle or how two points are close from each other. Some new application allows users to browse objects based on both of their geometric coordinates and their associated texts. Such type of queries called as spatial keyword query. For example, if a search engine can be used to find nearest hotels to a specific location that contain facilities lunch and dry cleaning at the same time. From this query, we could first obtain entire hotels whose facilities contain the set of keywords, and then find the nearest one from the retrieved hotels. The major drawback of this approach is that, on the difficult input they do not provide real time answer. For example, from the query point the real neighbor lies quite far away, while all the closer neighbors are missing at least one of the query keywords. In the past years, the group of people has showed interest in studying keyword search in relational databases. Recently the attention has preoccupied to multidimensional data (5)(6). The best method for nearest neighbor search with keywords is because of Felipe et al. (5).