Minimum description length pruning and maximum mutual information training of adaptive probabilistic neural networks

An approximated version of the minimum description length criterion (MDL) is applied to find optimal size adaptive probabilistic neural networks (APNNs) by adaptively pruning Gaussian windows from the probabilistic neural network (PNN). The authors discuss and compare both stochastic maximum likelihood (ML) and stochastic maximum mutual information (MMI) training applied to the APNN, for probability density estimation (PDF) and pattern recognition applications. Results on four benchmark problems show that the APNN performs better than or similar to the PNN, and that its size is optimal and much smaller than that of the PNN