Davies Bouldin Index Calculator

Work out davies bouldin index instantly with clear inputs, formula shown and shareable results.

About Davies Bouldin Index Calculator

The Davies-Bouldin index measures, for each pair of clusters, how large their internal spread is relative to the distance between their centres: R = (Si + Sj) / Dij. The overall index is the mean of each cluster's worst-case pairing, and lower is better. A value near or above 1 means the clusters are as spread out as they are far apart, so they overlap.

Formula

Pairwise similarity

R(i,j) = (Si + Sj) / D(i,j); DB index = mean over i of max over j of R(i,j)

Si: Average distance of cluster i's points to its centroid
D: Distance between centroids

Worked example

With the default values below, Davies Bouldin Index Calculator returns Davies-Bouldin similarity R of 0.6250. Change any input above and every figure here is recalculated in your browser.

Inputs used

Average scatter of cluster Aminimum 0.001: 0.9
Average scatter of cluster Bminimum 0.001: 1.1
Distance between centroidsminimum 0.001: 3.2
Clusters in the modelminimum 2: 2

Results

Davies-Bouldin similarity R: 0.6250
Combined within-cluster scatter: 2.0000
Centroid distance minus combined scatter: 1.2000
Reading: Acceptable separation across 2 clusters

Step by step

1Values used
2Pairwise similarity
3Davies-Bouldin similarity R= 0.6250
4Combined within-cluster scatter= 2.0000
5Centroid distance minus combined scatter= 1.2000
6Reading= Acceptable separation across 2 clusters

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to 4 decimal places, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

Is lower always better?

Lower is better within one dataset and one distance metric, but the index has no absolute scale, so use it to rank candidate clusterings rather than as a quality certificate.

How does it compare with the silhouette?

Davies-Bouldin only uses centroids and average scatter, so it is cheap but biased toward convex, equally sized clusters. The silhouette uses all pairwise distances and handles irregular shapes better.