Learning Vector Quantization.*;

Introduction

based on articles by Laurene Fausett, T. Kohonen, and Pandya & Macy.

"While VQ and the basic SOM are unsupervised clustering and learning methods, LVQ describes supervised learning. On the other hand, unlike in SOM, no neighborhoods around the "winner" are defined during learning in the basic LVQ...(Kohonen(2001), p. 245)".

You might have noticed that learning vector quantization is also covered in the Clustering area of this site. Here, the neural network version of learning vector quantization is a little different than the traditional version. The traditional version is based more on statistical pattern recognition, creating prototypes from the commonalities between various patterns (i.e.; attempting to find the clusters). The neural network version of LVQ, on the other hand, operates under the assumption that the centroids or prototypes are already known and that all input vectors are expected to fit into these clusters. Neural network LVQ is heavily dependent on supervised learning.

 

Works Cited:

Fausett, Laurene.  (1994).  Fundamentals of neural networks: Architectures, algorithms, and applications.  New Jersey: Prentice Hall.

Kohonen, Teuvo. (2001). Self-organizing maps. (3rd ed.). New York: Springer-Verlag Berlin Heidelberg.

Pandya, Abhijit S. & Macy, Robert B. (1995). Pattern recognition with neural networks in c++. Florida: CRC Press LLC.

 

public void footer() {
About | Contact | Privacy Policy | Terms of Service | Site Map
Copyright© 2009-2012 John McCullock. All Rights Reserved.
}