Competitive Learning Neural Networks (CLNNs) [Hertz et al., 1991] classify
unlabelled data sets, providing single layer cluster analysis. The nature of the
classification found by a CLNN is strongly influenced by the number of competitive
nodes used, which must be decided upon before starting. As CLNNs produce single
layer classifications they are not suitable for use in application areas such as biological
taxonomy [Sneath and Sokal 1973] that require a hierarchical analysis of the data.