Hierarchical Cluster Analysis

Hierarchical Clustering

  • Hierarchical cluster analysis in Aabel is comprised of agglomerative techniques for finding clusters of observations within a data set:
    • Objects/observations (worksheet row) are defined by a set of numeric variables; each object is positioned in a multi-dimensional space of a dimension proportional to the number of variables used to define the object.
  • This method requires n columns of continuous data.
The Agglomerative Models
  • Single Linkage (Nearest neighbor)
  • Complete Linkage (Furthest Neighbor)
  • McQuitty's Method (WPGMA)
  • UPGMA: Average Linkage Between Groups (Unweighted)
  • Gower's Method (Median)
  • Ward's Method
  • Centroid Method
The Dissimilarity or Distance Measure Options
  • Euclidean distance
  • Squared Euclidean distance
  • Manhattan distance
  • (1-r)/2 correlation coefficient dissimilarity
  • 1-|r| correlation coefficient dissimilarity
  • Bray-Curtis dissimilarity
  • Zero-adjusted Bray-Curtis dissimilarity
The Analysis Output

Hierarchical Cluster Analysis of a Published Dataset*

Hierarchical Cluster Analysis Dendrogarm

* Source of data: Naldrett, A.J., Hewins, R.H., Dressler, B.O., and Rao, B.V. (1984). The Contact Sublayer of the Sudbury Igneous Complex, in the Geology and Ore Deposits of the Sudbury Structure (Edited Pye, E.G., Naldrett, A.J., and Gilblin, P.E.), pp. 253-274, Ministry of Natural Resources, Ontario.