A new method of reducing network complexity in probabilistic neural network for target identification

The performance of Probabilistic Neural Network (PNN) in object identification has been investigated with reduced complexity retaining its remarkable recognition accuracy. A new method has been proposed to reduce the network complexity in PNN by performing the harmonics mean, geometry mean, arithmetic mean and root mean square values on the training samples. The simulation results have proven a significant improvement in network complexity reduction in PNN while maintaining the recognition accuracy. This would greatly enhance the applicability of PNNs.

[1]  Donald F. Specht,et al.  Probabilistic neural networks , 1990, Neural Networks.

[2]  Billur Barshan,et al.  Neural network based target differentiation using sonar for robotics applications , 2000, Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065).

[3]  D. F. Specht,et al.  Probabilistic neural networks for classification, mapping, or associative memory , 1988, IEEE 1988 International Conference on Neural Networks.

[4]  A. Vogt.  Should activation functions be affinely recursive? , 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.

[5]  M.V.C. Rao,et al.  Sequential learning neural network for sonar target differentiation , 2004, Proceedings. 2004 International Conference on Information and Communication Technologies: From Theory to Applications, 2004..

[6]  Richard P. Lippmann,et al.  An introduction to computing with neural nets , 1987 .

[7]  D. F. Specht,et al.  The use of probabilistic neural networks to improve solution times for hull-to-emitter correlation problems , 1989, International 1989 Joint Conference on Neural Networks.

[8]  Donald F. Specht,et al.  Applications of probabilistic neural networks , 1990, Defense, Security, and Sensing.