Title :
An analytical alternative for SOM
Author :
Menhaj, Mohammad B. ; Jahanian, H.R.
Author_Institution :
Electr. Eng. Dept., Amirkabir Univ. of Technol., Tehran, Iran
Abstract :
This paper introduces a new self-organizing system, called a continuous self organising map (CSOM). The behavior of this model, which is represented by a set of nonlinear differential equations, is identical to Kohonen´s self-organizing neural network. This paper shows that CSOM can be viewed as an appropriate non-algorithmic model for the class of one-dimensional Kohonen algorithms with an arbitrary number of inputs
Keywords :
generalisation (artificial intelligence); learning (artificial intelligence); nonlinear differential equations; self-organising feature maps; topology; 1D topology; Kohonen SOM; continuous self organising map; generalisation; learning; neural network; nonlinear differential equations; Artificial neural networks; Biological system modeling; Closed-form solution; Differential equations; Network topology; Neural networks; Nonlinear equations; Probability distribution; Prototypes;
Conference_Titel :
Neural Networks, 1999. IJCNN '99. International Joint Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5529-6
DOI :
10.1109/IJCNN.1999.832679
