ML Wiki

DBSCAN

It's a density-based clustering algorithm

Density associated with a point is obtained by counting the number of points in a region of specified radius $\epsilon$ around each point

• points with density $\geqslant \text{min_pts}$ are considered as "core points"
• noise and non-core points are discarded
• clusters are formed around the core points
• if two core points are within a radius $\epsilon$, then they belong to the same cluster

Disadvantages

• can find clusters of different shapes, but can't find clusters of different densities

Extensions

SNN Clustering

• an extension of DBSCAN that words better for high-dimensional data
• also can find clusters of different density

References

• Ester, Martin, et al. "A density-based algorithm for discovering clusters in large spatial databases with noise." 1996. [1]

Sources

• Ertöz, Levent, Michael Steinbach, and Vipin Kumar. "Finding clusters of different sizes, shapes, and densities in noisy, high dimensional data." 2003. [2]