Cell-Splitting Grid : A Self-Creating and Self-Organizing Neural Network

A new model of self-creating and self-organizing neural network, Cell-Splitting Grid (CSG), is presented. In this proposed CSG algorithm, the neurons and their connections are created and organized on a 2-dimensional plane according to the input data distribution. Compared with Self-Organizing Map (SOM) and its other improved algorithms, the CSG algorithm has outperformed the SOM and other SOM related algorithms in vector quantization while maintaining relatively good topology preservation. This paper shows that CSG is a promising and an effective method especially for non-uniformly distributed data