Understanding OPTICS Clustering: Hands-On With Scikit-Learn
Unsupervised Learning — Clustering
OPTICS → Ordering Points To Identify the Clustering Structure
The OPTICS algorithm is similar to the DBSCAN algorithm, but it allows for a range of values for the ‘eps’ parameter instead of just one.
The OPTICS algorithm creates a reachability plot (we will learn more about it below). If we run the OPTICS algorithm setting the values for ‘eps’ and ‘min_points’ we can transform the OPTICS in a DBSCAN algorithm. For now, let’s learn more about the OPTICS parameters:
Epsilon → Is a distance around a point. It is not a distance o any other point, but rather a line that defines a circumference (area) around a point.
Minimum points → Defines the minimum number of points that must be within the epsilon distance.
Core points → We can say a given point is a core point when that point has the minimum number of points within its epsilon distance. A core point represents a cluster.
