Engineering applications of the self-organizing map

TLDR

The self‑organizing map (SOM) is a powerful, unsupervised, error‑theoretic tool that visualizes and compresses high‑dimensional data while preserving topological and metric relationships. It transforms complex, nonlinear statistical relationships into simple geometric relationships on a low‑dimensional display. Industries report that SOM’s orderly input‑output mapping reduces training data needs, accelerates nonlinear interpolation and extrapolation, enhances generalization, and enables efficient information compression.

Abstract

The self-organizing map (SOM) method is a new, powerful software tool for the visualization of high-dimensional data. It converts complex, nonlinear statistical relationships between high-dimensional data into simple geometric relationships on a low-dimensional display. As it thereby compresses information while preserving the most important topological and metric relationships of the primary data elements on the display, it may also be thought to produce some kind of abstractions. The term self-organizing map signifies a class of mappings defined by error-theoretic considerations. In practice they result in certain unsupervised, competitive learning processes, computed by simple-looking SOM algorithms. Many industries have found the SOM-based software tools useful. The most important property of the SOM, orderliness of the input-output mapping, can be utilized for many tasks: reduction of the amount of training data, speeding up learning nonlinear interpolation and extrapolation, generalization, and effective compression of information for its transmission.