基本概念:(Density-Based Spatial Clustering of Application with Noiso)
1. Core Objects:
If a point density reaches a threshold algorithm set it as the core point. (I.e., the number of points in the field of not less than MinPts r)
FIELD 2.ε- distance threshold:
Set radius r
3. Direct density of up to:
If a point p in the field of point q r, and q is the core points density-reachable directly pq
4. density of up to:
If a point sequence q0, Ql, ... qk, for any qi-qi-q are density-reachable directly, called up from q0 to qk density, the density of which is in fact directly reachable "communication . "
5. Density is connected:
If starting from a core point p, q point and k are density-reachable point, called the point q and k is the density of points connected
6. boundary points:
Non-core point belong to a class, we can not develop off the assembly line
7. Direct density of up to:
If a point p in the field of point q r, and q is the core points density-reachable directly pq
8. Noise Point:
Does not belong to any class cluster of points, starting from any point is a core density is not reachable
9. The visual display:
A: Core Objects
B, C: boundary point
N: outlier