Title :
A training scheme for pattern classification using multi-layer feed-forward neural networks
Author :
Keeni, Kanad ; Nakayama, Kenji ; Shimodaira, Hiroshi
Author_Institution :
Dept. of Inf. Syst. & Quantitative Sci., Nanzan Univ., Japan
Abstract :
This study highlights the subject of weight initialization in multi-layer feed-forward networks. Training data is analyzed and the notion of critical point is introduced for determining the initial weights for input to hidden layer synaptic connections. The proposed method has been applied to artificial data. Experimental results show that the proposed method takes almost half the training time required for standard backpropagation
Keywords :
backpropagation; feedforward neural nets; multilayer perceptrons; pattern classification; backpropagation; critical point; hidden layer; initial weights; multilayer feedforward neural networks; pattern classification; synaptic connections; training data; training scheme; weight initialization; Feedforward neural networks; Feedforward systems; Multi-layer neural network; Neural networks; Pattern classification;
Conference_Titel :
Computational Intelligence and Multimedia Applications, 1999. ICCIMA '99. Proceedings. Third International Conference on
Conference_Location :
New Delhi
Print_ISBN :
0-7695-0300-4
DOI :
10.1109/ICCIMA.1999.798548
