A fast OBS pruning algorithm based on pseudo-entropy of weights

A fast OBS pruning algorithm based on pseudo-entropy of weights is proposed to resolve the problems of the number of hidden neurons is difficult to be determined in neural networks and low pruning speed in conventional OBS (Optimal Brain Surgeon) pruning algorithm. The algorithm makes the network constrain the distribution of weight automatically during the training process, obtain a simpler structure of network, and improve the speed of pruning. The result of experimental shows that the structure of network has been simplified; the generalization capability of network and the speed of pruning have been improved greatly by using the algorithm mentioned above.