DBSCAN (Density-based spatial clustering of applications with noise)
As opposed to k-means, the DBSCAN algorithm does not require a user-defined k (number of clusters). It assigns densely-connected sample points to the same cluster, putting aside the less densely regions within the data points. As such it is robust to the presence of outliers and is able to form clusters of arbitrary shape. It is however non-deterministic ref and its performance degrades at high dimensions ref.
Applications
- Single neuronal methylomes obtained from mice and human frontal cortices were compared using DBSCAN ref
- Edge detection of skin lesions obtained from dermoscopic images ref
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.