AMORE (version 0.1.1)

train.compare: Trains the same neural network according to different error criteria.

Description

This function trains a neural network according to different error criteria so as to compare the different behaviours.

Usage

train.compare(net.start, P, T, ideal=NA, max.epoch, show.step, Stao=1000, criteria=c("MSE","LMLS","TAO"))

Arguments

net.start

Neural Network to train.

P

Training set input values.

T

Training set output values

ideal

Clean training set output values. Useful for training robust networks with noisy data.

max.epoch

Number of epochs to train.

show.step

A report is provided every show.step epochs.

Stao

Initial value of the S parameter used by the TAO algorithm.

criteria

A vector specifying which criteria should be used.

Value

• This function returns a list containing the trained Neural Network objects according to the specified criteria.

encoding

latin1

References

Pernia Espinoza, A.V. TAO-robust backpropagation learning algorithm. Neural Networks. In press. Simon Haykin. Neural Networks. A comprehensive foundation. 2nd Edition.

See Also

train