Do you have any idea about wavelet neural networks? What is the influence of changing neural network's transfer function with mother wavelet functions?
The wavelet neural network is otherwise referred as Wavenets. By changing the T.F using wavelet functions, the network will try to reduce the error factor(if any) in the inputs and processes/approximates accordingly to get the desired output. It is expected that the Wavenets will provide good results based on the performance analysis. Moreover it is better fit for problems with nonlinearity and non stationary data's.
These networks (wavelets networks) allow the representation of a non linear function by comparing their inputs and their outputs while training. This training is made while representing a non linear function by a combination of activation functions. The sigmoid function is often used as an activation one. The input of this prototype is a set of parameters. So the entries are not actual data but only values describing specific positions of the analyzed signal. The hidden layer contains a set of nodes; each node is composed by a translated and dilated wavelet. The output layer contains one node which sums the outputs of the hidden layer by weighted connections weights.
- Badges
- Science topic