A static R-Tree organization method based on top-down recursive clustering

Traditional R-Tree organization methods usually choose a bottom-up method, while a small number of top-down algorithms cannot give an optimal performance due to the limitations of the division mode. This paper proposed a new R-tree organization method based on top-down recursive clustering against the deficiencies of the traditional bottom-up divided structure node. This method divided our spatial data into several classes by a K-means clustering method and uses the method of STLT to build the R-tree top-down, instead of adjusting each class. Apply this method to all the classes by recursions to construct the whole tree. The experiment shows that the new algorithm gives a better performance in query efficiency than Hilbert and STR, but has a bad construction efficiency to be improved.