Figure 2
Graphical representation of the spatial density-based clustering results produced by OPTICS (Ankerst et al., 1999 ). Data points are represented as black dots, with circles of different colours (A, B, C and D) representing clusters. By not including a global density requirement, clusters of different spatial densities can be automatically separated (A, B, C and D), whereas a global density requirement would either join C and D together or omit A and B entirely. The same concept applies in higher dimensions. |