The Kohonen SOM falls into the category of unsuper-vised learning method, in which the relevant multivariate algorithms seek clusters in the data. The SOM was proposed by Kohonen in the early 1980s (1982, 1984). Since that time, the SOM has been used in a number of different applications in diverse field and it has been the most well-known ANN with unsupervised learning rules. The algorithm performs a topology-preserving projection of the data space onto a regular low-dimensional space (usually a two-dimensional space) and can be used to effectively visualize clusters. It is widely applicable to the fields of data mining, data classification, and biological modeling in terms of a nonlinear projection of multivariate data into lower dimensions.
Was this article helpful?