Cluster Analysis using Spherical SOM
Heizō Tokutaka, Kihato, P. K., Fujimura, Kikuo, Ohkita, Masaaki · 2019
A cluster analysis method is proposed in this paper. As benchmark data, the Fisher's iris and the Wine recognition data sets are used. As a result of the numerical experiment, a clustering method using the dendrogram yielded 97 % in accuracy. It is difficult to display a multi-dimensional data by the dendrogram in one dimension. The ultimate visualization is by means of 3 dimensional rendition. We conclude that the best way that a multi-dimensional data set is visualized is by a sphere, since the phase relationship of it is smooth everywhere.