The Davies–Bouldin Index (or DBI) is a measure that is used to assess how well clusters in a cluster analysis can separate data points that are different from each other. It works by comparing the distance between two clusters to the amount of variation (or 'spread') within each cluster. If a cluster with a small amount of variation is far from another cluster with a large amount of variation, it means that it is doing a good job of separating the data points in the clusters.