Canopy clustering algorithm
The algorithm proceeds as follows:
- Cheaply partitioning the data into overlapping subsets (called "canopies")
- Perform more expensive clustering, but only within these canopies
Since the algorithm uses distance functions and requires the specification of distance thresholds, its applicability for high-dimensional data is limited by the curse of dimensionality. Only when a cheap and approximative - low dimensional - distance function is available, the produced canopies will preserve the clusters produced by K-means.
The method first appeared in a paper by Andrew McCallum, Kamal Nigam and Lyle Ungar.
- The number of instances of training data that must be compared at each step is reduced
- There is some evidence that the resulting clusters are improved
- Mahout description of Canopy-Clustering Retrieved 2011-04-02.
- McCallum, A.; Nigam, K.; and Ungar L.H. (2000) "Efficient Clustering of High Dimensional Data Sets with Application to Reference Matching", Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining, 169-178 doi:10.1145/347090.347123
|This algorithms or data structures-related article is a stub. You can help Wikipedia by expanding it.|