Interpretation of a Dendrogram

Since we have had a couple of blogs on clustering, I would like to talk about the representation of the clustering analysis. A Dendrogram is a visual representation of the correlation between various clusters. This diagram shows at what level of similarity the clusters were joined. This is indicated by the distance shown on the y-axis. The point at which the clusters are joined is known as the node.

Below is an example of a dendrogram.

The purpose of a dendrogram is usually to determine the number of distinct clusters that can be formed. With repeated clustering, a single cluster is formed. The distance from one cluster from another indicates the degree of relationship, i.e closely related clusters are placed close to one another, the closeness of which can be determined on the y-axis.

Whenever there is a leap in the distance between two nodes, we can break them into a different cluster. This process is subjective and there is no limit on the number of clusters that can be made.

Below are the links that helped me write this blog.

http://www.nonlinear.com/support/progenesis/samespots/faq/dendrogram.aspx

http://strata.uga.edu/software/pdf/clusterTutorial.pdf