How Many Dissimilarity/Kernel Self Organizing Map Variants Do We Need?

TL;DR

This paper reviews various dissimilarity and kernel-based Self Organizing Map variants, comparing their differences, advantages, drawbacks, and practical relevance in handling complex data beyond numerical vectors.

Contribution

It provides a unified comparison of SOM variants based on dissimilarity and kernel methods, highlighting their differences and practical applicability.

Findings

Different SOM variants have distinct advantages and limitations.

The relevance of dissimilarity/kernel SOM varies across applications.

A unified notation helps compare and understand SOM variants.

Abstract

In numerous applicative contexts, data are too rich and too complex to be represented by numerical vectors. A general approach to extend machine learning and data mining techniques to such data is to really on a dissimilarity or on a kernel that measures how different or similar two objects are. This approach has been used to define several variants of the Self Organizing Map (SOM). This paper reviews those variants in using a common set of notations in order to outline differences and similarities between them. It discusses the advantages and drawbacks of the variants, as well as the actual relevance of the dissimilarity/kernel SOM for practical applications.