Self-Learning and Adaptive Algorithms for Business Applications by Hu Zhengbing;Bodyanskiy Yevgeniy V.;Tyshchenko Oleksii;

Author:Hu, Zhengbing;Bodyanskiy, Yevgeniy V.;Tyshchenko, Oleksii;
Language: eng
Format: epub
Publisher: Emerald Publishing Limited
Published: 2019-05-23T00:00:00+00:00

3.2. KOHONEN NEURAL NETWORKS

Self-organizing maps (SOMs (Kohonen, 2001)) were initially designated for reduction of data dimensionality, and are typically used for visualization tasks, data compression, modeling, and so on. An SOM’s peculiarity is the fact that some additional non-adjustable vector parameters are used (in addition to using a matrix of weights ) where each parameter is an image of a corresponding node (a column vector of the weights’ matrix ) into the space of reduced dimensionality. Nodes’ images are arranged into a regular lattice, and its topology can be different but is usually rectangular or hexagonal (Figure 3.2).

Figure 3.2. Common Types of SOM Topologies: (а) Rectangular; (b) Hexagonal.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.