Density-Based Spatial Clustering of Applications with Noise

Density-based spatial clustering of applications with noise (DBSCAN) is a data clustering algorithm proposed in [Ester96]. It is a density-based clustering non-parametric algorithm: given a set of observations in some space, it groups together observations that are closely packed together (observations with many nearby neighbors), marking as outliers observations that lie alone in low-density regions (whose nearest neighbors are too far away).

Details

Given the set $LaTex Math image.$ of n
p
-dimensional feature vectors (further referred as observations), a positive floating-point number

epsilon

and a positive integer

minObservations

, the problem is to get clustering assignments for each input observation, based on the definitions below [Ester96]:

core observation: An observation x
is called core observation if at least
minObservations
input observations (including x
) are within distance
epsilon
from observation x
;
directly reachable: An observation y
is directly reachable from x
if y
is within distance
epsilon
from core observation x
. Observations are only said to be directly reachable from core observations.
reachable: An observation y
is reachable from an observation x
if there is a path $LaTex Math image.$ with $LaTex Math image.$ and $LaTex Math image.$ , where each $LaTex Math image.$ is directly reachable from $LaTex Math image.$ . This implies that all observations on the path must be core observations, with the possible exception of y
.
noise observation: Noise observations are observations that are not reachable from any other observation.
cluster: Two observations x
and y
are considered to be in the same cluster if there is a core observation z
, and x
and y
are both reachable from z
.

Each cluster gets a unique identifier, an integer number from 0
to $LaTex Math image.$ . Each observation is assigned an identifier of the cluster it belongs to, or -1
if the observation considered to be a noise observation.