Recently, "spatial data" bases have been extensively adopted in the recent decade and various methods have been presented to store, browse, search and retrieve spatial objects". In this study, a method is plotted for retrieving nearest neighbors from "spatial data" indexed by R+tree. The approach uses a reduced R+tree for the purpose of representing the "spatial data". Initially the "spatial data" is selected and R+tree is constructed accordingly. Then a function called joining nodes is applied to reduce the number of nodes by combining the half-filled nodes to form completely filled. The idea behind reducing the nodes is to perform search and retrieval quickly and efficiently. The reduced R+tree is then processed with KNN query algorithm to fetch the nearest neighbors to a point query. The basic procedures of KNN algorithm are used in the proposed approach for retrieving the nearest neighbors. The proposed approach is evaluated for its performance with "spatial data" and results are plotted in the experimental analysis section. The experimental results showed that the proposed approach is remarkably up a head than the conventional methods. The maximum time required to index the 1000 data points by the R+tree is 10324 ms. The number of nodes possessed by reduced R+tree is also less for 1000 data points as compared to the conventional R+tree algorithm.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。