Abstract | ||
---|---|---|
In order to better construct the R - tree index structure and improve the efficiency of R - tree construction and retrieval, the R - tree construction algorithm is optimized from the following three aspects: the selection and determination of the initial center point, the weighted distance between data combined with the actual situation, and the division of redundant data by combining the theories of nearest neighbor, information entropy and probability statistics. The feasibility of the proposed algorithm is verified by comparing the objective function values, iteration times and time consumption under different cluster numbers. Furthermore, experimental results show that the index structure constructed by this method has smaller overlapping area and more efficient query efficiency. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/FSKD.2018.8687189 | ICNC-FSKD |
Field | DocType | Citations |
k-nearest neighbors algorithm,k-means clustering,R-tree,Probability and statistics,Mathematical optimization,Weighted distance,Computer science,Algorithm,Entropy (information theory),Spatial database | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 5 |