Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 

DBSCAN in kdb+/q

DBSCAN (Density Based Spatial Clustering of Applications with Noise) is a clustering technique introduced at KDD-96.

A cluster is defined to be formed by "core" points and "border" points:

  • Core points have some minimum number of points in its neighborhood
  • Border points have at least one core point in its neighborhood

The approach to DBSCAN presented here augments the original algorithm by relying on a Euclidean distance matrix to find all neighbor points.

Then we use the Disjoint-set data structure to form the clusters by doing a union on pairs of collocated points.

Unfortunately this approach does not handle more than a few thousand points and only supports a Euclidean distance metric. Future work would address these issues.