Optimal Training Sequences for Locally Recurrent Neural Networks
The problem of determining an optimal training schedule for a locally recurrent neural network is discussed. Specifically, the proper choice of the most informative measurement data guaranteeing the reliable prediction of the neural network response is considered. Based on a scalar measure of the performance defined on the Fisher information matrix related to the network parameters, the problem was formulated in terms of optimal experimental design. Then, its solution can be readily achieved via the adaptation of effective numerical algorithms based on the convex optimization theory. Finally, some illustrative experiments are provided to verify the presented approach.
KeywordsInput Sequence Recurrent Neural Network Fisher Information Matrix Optimum Experimental Design Experimental Design Theory
Unable to display preview. Download preview PDF.
- 4.Patan, K.: Artificial Neural Networks for the Modelling and Fault Diagnosis of Technical Processes. LNCIS. Springer, Berlin (2008)Google Scholar
- 7.Patan, K., Patan, M.: Selection of training sequences for locally recurrent neural network training. In: Malinowski, K., Rutkowski, L. (eds.) Recent Advances in Control and Automation, pp. 252–262. Academic Publishing House, EXIT, Warsaw (2008)Google Scholar