Search

Proposes a new competitive learning algorithm that dynamically creates output neurons. The number of output neurons is increased as learning proceeds, whereas conventional competitive learning algorithms use all of the available output neurons during the entire learning phase. An acceptance test for the winning output neuron is performed using class thresholds if the number of created output neurons is less than the predefined maximum number. Accepted input vectors are used to adjust the reference vector of the winning output neuron. If an input vector is rejected, it is used as the initial reference vector of a new output neuron. The proposed method gets around the drawbacks of the conventional competitive learning algorithms by changing the class threshold values of output neurons dynamically. Experiments with remote sensing data and speech data indicate the superiority of the proposed algorithm in comparison to the conventional competitive learning methods. >

In this paper, we propose a new multiple self-organizing neural network architecture. Whereas conventional learning methods utilize the full dimension of the original input patterns, the proposed system consists of multiple neural networks with the reduced input dimension. We have developed three consensus schemes so as to judge the classification using multiple neural networks. Each network of ensemble has dynamic properties. The number of output neurons is increased as learning proceeds. Every output neuron has its own class threshold, which represents a class boundary. The class threshold value is tuned according to input pattern distribution. The performance of the multiple self organizing neural network is compared with those of conventional competitive learning algorithms.