FAULT-TOLERANT TRAINING FOR OPTIMAL INTERPOLATIVE NETS

The optimal interpolative (OI) classification network is extended to i nclude fault tolerance and make the network more robust to the loss of a neuron. The OI net has the characteristic that the training data ar e fit with no more neurons than necessary. Fault tolerance further red uces the number of neurons generated during the learning procedure whi le maintaining the generalization capabilities of the network, The lea rning algorithm for the fault-tolerant OI net is presented in a recurs ive format, allowing for relatively short training times, A simulated fault-tolerant OI net is tested on a navigation satellite selection pr oblem.