Davies-Bouldin Lite

Problem

Compute the Davies-Bouldin index for exactly two clusters: (spread_0 + spread_1) / distance(centroid_0, centroid_1), where spread is the mean Euclidean distance of cluster members to their centroid.

Return 0.0 if there are not exactly two clusters, the centroids coincide, or input is empty.

Hint

Compact, well-separated clusters give a small Davies-Bouldin index.

Test cases

case 1

in: [[[0,0],[1,0],[10,0],[11,0]],[0,0,1,1]]

expected: 0.1

case 2

in: [[[0],[1],[2],[10]],[0,0,0,1]]

expected: 0.07407407407407407

case 3

in: [[[0,0],[1,1]],[0,0]]

expected: 0

case 4

in: [[],[]]

expected: 0

Starter codePython

def davies_bouldin_2(points, labels):
    # Your code here
    pass

4 test cases · run your code against them in the editor