Improved interpolation and extrapolation from continuous training examples using a new neuronal model with an adaptive steepness
David McLean, Zuhair A. Bandar, James D. O’Shea · 2002
The ability of a neural network to generalise is dependant on how representative the training patterns were, of the whole data domain, and how smoothly the network has fitted to these patterns. In non-scaled continuous data domains training examples will lie at differing distances from each other, making the fitting problem more difficult and varied. This paper introduces a new fitting technique in which each neuron is given an adaptive steepness parameter, implemented as an extra internal connection, which is altered to better interpolate between the data points that its hyperplane divides. An example benchmark problem is used to illustrate the effects of this technique and results for a real world data domain are given which display improved an classification rate when compared against networks with a constant steepness value for every neuron.>