ABSTRACT
In Chapters 5 and 6, we presented various methods for clustering, including agglomerative clustering, k-means clustering, and model-based clustering. In the process of doing that, we showed some visualization techniques such as dendrograms for visualizing hierarchical structures and scatterplots. We now turn our attention to other methods that can be used for visualizing the results of clustering. These include a space-filling version of dendrograms called treemaps, an extension of treemaps called rectangle plots, a new rectangle-based method for visualizing nonhierarchical clustering called ReClus, and data images that can be used for viewing clusters, as well as outlier detection. We begin by providing more information on dendrograms.
The dendrogram (also called the tree diagram) is a mathematical, as well as a visual representation of a hierarchical procedure, which can be divisive or agglomerative. Thus, we often refer to the results of the hierarchical clustering as the dendrogram itself.
