Using the important property of the approximating a posteriori probability functions of the classes in the outputs of the trained multilayer perceptrons, we propose the technique for the implementation of sequential classification by perceptron and multilayer perceptron, and application to the node growing in the number of input nodes of perceptron and the number of hidden nodes of the multilayer perceptron. A measurement for the ordering of hidden nodes of the trained multilayer perceptron is also proposed. The ordering of the hidden nodes comes from the contribution of each hidden node. Using the node growing technique, the minimum number of hidden nodes can be obtained in the training and used in the classification. The technique can also apply to the single layer perceptron. In the experiment, the typical 'XOR' problem is applied. The balance between the reduction of hidden nodes and classification results is quite good.
|Number of pages||6|
|State||Published - 1 Dec 1994|
|Event||Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7) - Orlando, FL, USA|
Duration: 27 Jun 1994 → 29 Jun 1994
|Conference||Proceedings of the 1994 IEEE International Conference on Neural Networks. Part 1 (of 7)|
|City||Orlando, FL, USA|
|Period||27/06/94 → 29/06/94|