This book is Work in Progress. I appreciate your feedback to make the book better.

8.2 Clusters

Cluster analysis

Cluster analysis searches for groups of similar observations. The target is usually a partition of persons, schools, countries, documents, or other units rather than a continuous latent dimension.

A cluster can be interpreted descriptively without claiming that it is a naturally existing type. Later extensions such as latent class and latent profile models make the grouping probabilistic and connect clustering to the wider family of latent-variable models.