A compensatory wavelet neuron model

The paper proposes a compensatory wavelet neural model, which is based on a wavelet activation function. The basis function comprises both summation and multiplicative functions. It is shown by M. Sinha et al. that, for a spectrum of functional mapping and classification problems, the compensatory neuron based neural network model performs better than the ordinary neuron based neural network, in terms of both the accuracy of prediction and the computational time involved. On the other hand, the wavelet neuron is obtained by modifying an ordinary neuron with non-orthogonal wavelet bases (T. Yamakawa et al., 1994). The performances of different neuron based neural networks are also analyzed