Density-Based Spatial Clustering of Applications with Noise is a clustering algorithm able to find non-linearly separable and arbitrarily-shaped clusters given a radius and density constraint. In addition, DBSCAN also has the ability to mark outliers as noise and thus can be used as a quasi anomaly detector.
Note: Noise samples are assigned the cluster number -1.
Data Type Compatibility: Continuous
|1||radius||0.5||float||The maximum distance between two points to be considered neighbors.|
|2||min density||5||int||The minimum number of points within radius of each other to form a cluster.|
|3||tree||BallTree||object||The spatial tree used to run range searches.|
This estimator does not have any additional methods.
use Rubix\ML\Clusterers\DBSCAN; use Rubix\ML\Graph\Trees\BallTree; use Rubix\ML\Kernels\Distance\Diagonal; $estimator = new DBSCAN(4.0, 5, new BallTree(20, new Diagonal()));
- M. Ester et al. (1996). A Densty-Based Algorithm for Discovering Clusters.