How do I interpret error metrics for my clustering model?

Once a clustering model is trained the following error metrics are generated:

  • Silhouette Score: the Silhouette score is a commonly used measure of how well separated and cohesive the resulting clusters are. Note that the Silhouette score is most relevant when dealing with well-behaved, i.e. convex-shaped, clusters. In many real-life cases clusters are not convex and the Silhouette score may not be as relevant. In this case you are best served relying on your domain expertise to identify the most relevant number of clusters. 
  • Computed Clusters: the number of clusters generated by the model. It will generally be the number of components specified during model configuration unless you enabled optimization, in which case it is the computed number of clusters that maximizes the Silhouette score. 

Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.