Learning vector quantization: Difference between revisions

Revision as of 17:20, 5 May 2023

Revision as of 09:49, 18 September 2023

Line 2:

== Overview ==

LVQ can be understood as a special case of an [[artificial neural network]], more precisely, it applies a [[winner-take-all (computing)|winner-take-all]] [[Hebbian learning]]-based approach. It is a precursor to [[self-organizing map]]s (SOM) and related to [[neural gas]], and to the [[k-nearest neighbor algorithm]] (k-NN). LVQ was invented by [[Teuvo Kohonen]].<ref>T. Kohonen. Self-Organizing Maps. Springer, Berlin, 1997.</ref>

An LVQ system is represented by prototypes <math>W=(w(i),...,w(n))</math> which are defined in the [[feature space]] of observed data. In winner-take-all training algorithms one determines, for each data point, the prototype which is closest to the input according to a given distance measure. The position of this so-called winner prototype is then adapted, i.e. the winner is moved closer if it correctly classifies the data point or moved away if it classifies the data point incorrectly.