Cluster Analysis of Subjects, Hierarchical Methods
Barry J. Everitt · Encyclopedia of Biostatistics · 2005
Abstract Hierarchical cluster classification involves partitioning data into a series of groups. The first group consists ofnsingle‐member “clusters”, the last consists of a single group with allnindividuals. Hierarchical cluster classification could be represented by a diagram known as a dendrogram. Properties and problems of hierarchical clustering techniques are described. The two major types of algorithms that have been used to produce hierarchical classifications are agglomerative and divisive. Careful validation of solutions is a clear requirement in any clustering exercise. An example to illustrate hierarchical methods is presented.