DBSCAN

(Master of Information Science) Machine Learning

• If minPts points are inside a circle of radius ε, they are considered to belong to the same cluster.
  • If a point does not belong to any circle, it is considered as noise (not part of any cluster).
  • It is necessary to check if all points are inside the circle for each point.
    • Therefore, the time complexity is O(n^2) if done directly.
• Method for speeding up the process:
  • Discretize the points and use the “neighboring points” instead of the points inside the circle.
• Although it is an old technique, it is still widely used today. #Clustering