Overview

Some clustering methods use the distance between clusters as an element of the clustering algorithm. Because clusters contain multiple points, it is somewhat hard to define a single definition of distance.

Distance Measures

Given a metric and two sets {% C_a %} and {% C_b %}, the following distances can be defined

• Minimum Distance
  {% d(C_a,C_b) = min \{ d(x_i, x_j) ; x_i \in C_a \& x_j \in C_b \} %}
• Maximum Distance
  {% d(C_a,C_b) = max \{ d(x_i, x_j) ; x_i \in C_a \& x_j \in C_b \} %}
• Average Distance
• Distance between Centroids