Chung, F.L., and Lee, T. 1992. A node pruning algorithm for backpropagation networks. Int. J. of Neural Systems, 3(3), 301--314. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
A Penalty-Function Approach for Pruning Feedforward Neural Networks - Setiono (1994) (4 citations) (Correct)
....algorithm (Frean 1990) The aim of these algorithms is to find a network with the simplest architecture possible that is capable of correctly classifying all its input patterns. Interests in algorithms that remove hidden units from an oversized network have also been growing (Chauvin 1989; Chung and Lee 1992; Mozer and Smolensky 1989; Hanson and Pratt 1989) Instead of removing unnecessary hidden units, individual weights (connections) in the network may also be removed to increase the generalization capability of a neural network (Le Cun et al. 1990; Hassibi and Stork 1993; Thodberg 1991) ....
....Table 2: The average number of connections and hidden units obtained from 50 networks for the 4 bit and 5 bit parity problems after pruning. to solve the n bit parity problem. Previously published neural network pruning algorithms have also failed to obtain networks with less than n hidden units (Chung and Lee 1992; Hanson and Pratt 1989) In fact, 3 hidden units are sufficient for both the 4 bit and 5 bit parity problems. 3.3 The monks problems The monks problems (Thrun et al. 1991) are an artificial robot domain, in which robots are described by six different attributes: 19 A 1 : head shape 2 round, ....
Chung, F.L., and Lee, T. 1992. A node pruning algorithm for backpropagation networks. Int. J. of Neural Systems, 3(3), 301--314.
Extracting Rules From Pruned Neural Networks for Breast Cancer.. - Setiono (1996) (15 citations) (Correct)
....pattern will be classified as a member of a certain class. The complexity of a network can be reduced by removing its connections that are redundant through pruning. Many algorithms for neural network pruning have been proposed in the past few years. It has often been mentioned in the literature [6, 7, 8, 10, 17] that neural network pruning is beneficial in two ways. The first advantage is that a pruned network can achieve a higher accuracy rate on new patterns not used for training. The second advantage, which will be the focus of this paper, is that rules may be extracted from a network with a small ....
F.L. Chung and L. Lee, A node pruning algorithm for backpropagation network, Int. Journal of Neural Systems 3 (3) (1992) 301-314.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC