In the paper, a new kind of clustering algorithm called GCARD is proposed. Besides the merits of Density-Based clustering analysis and its efficiency, GCARD can capture the shape and extent of clusters by core grid units, and then analyze data based on the references of core grid units. We present a method of RGUBR to improve the accuracy of grid clustering method, so it can be used to discover information in very large databases.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
- A Grid Clustering Algorithm Based on Reference and Density
- Springer Berlin Heidelberg