Title :
Design artificial neural networks based on the principle of divide-and-conquer
Author_Institution :
Sch. of Comput. Sci., Tech. Univ. of Nova Scotia, Halifax, NS, Canada
Abstract :
A split-and-merge growth learning algorithm that builds a feedforward network of threshold logic units with two hidden layers to classify any given (nonlinearly) separable training patterns is proposed. The algorithm simultaneously learns the structure and connection weights of the network. Fast convergence in finite steps to the correct classification is guaranteed. It is based on the principle of divide-and-conquer. The ways in which this algorithm differs from other growth algorithms is discussed
Keywords :
learning systems; neural nets; artificial neural networks; classification; connection weights; divide-and-conquer; feedforward network; growth algorithms; hidden layers; split-and-merge growth learning algorithm; threshold logic units; training patterns; Artificial neural networks; Carbon capture and storage; Computational complexity; Computer networks; Computer science; Convergence; Logic; Merging; Partitioning algorithms; Polynomials;
Conference_Titel :
Circuits and Systems, 1991., IEEE International Sympoisum on
Print_ISBN :
0-7803-0050-5
DOI :
10.1109/ISCAS.1991.176614
