Sensitivity Analysis of Artificial Neural Networks Case Study in Clinical Electromyography

The usefulness of artificial neural networks (ANN) trained with the momentum back propagation (MBP) and the conjugate gradient back propagation (CGBP) learning algorithms in the classification of EMG data has recenlty been demonstrated. In this papes, we examine the sensitivity of feed forward layered networks supplied with EMG data and trained with the CGBP learning algorithm to i) weight errors and ii) random cut off of connections. Our results suggest that ANN models are capable of tolerating weight error changes around the optimal values, as well as limited number of disconnections