Numerical analysis of the RBF networks near singularities

Analyzing the learning dynamics near singularities of the feedforward neural networks is a research hotspot in recent years. It is meaningful to take a numerical analysis of the learning dynamics near the singularities in RBF networks. In this paper, we give the explicit expression of the Fisher information matrix, then we take a large number of simulation experiments to investigate the dynamics of learning of RBF networks. The simulation results indicate that the learning process is affected by the singularities seriously.