GLDBSCAN:An Improved Fast Clustering Algorithm
2009
A fast algorithm GLDBSCAN is proposed in this paper,which considerably speeds up the original LDBSCAN.The new algorithm presents a novel approach to grid generation and adopts the spatial index structure SP-Tree to organize the grids.In addition,GLDBSCAN chooses the representative object in a grid to represent all objects of the grid to cluster,which reduces the execution frequency and I/O cost dramatically.The experimental results show that GLDBSCAN is much more efficient than LDBSCAN without reducing the quality of clustering results and it is effective and efficient in clustering large-scale datasets.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI