nnetTrainSetup: nnetTrainSetup returns a list of templates for use by forward...
In philipmgoddard/nnePtR: Neural Network Classifier

nnetTrainSetup returns a list of templates for use by forward and back propogation functions. Far more efficient to pass templates (defined once) than redefining matrices of correct size at each iteration when optimising

1
2
3

nnetTrainSetup(input, outcome, nLayers = 1, nUnits = 10, seed = 1234)

nnetTrainSetup_c(input, outcome, nLayers = 1, nUnits = 10, seed = 1234)

`input`	matrix of inputs (nSamples x nFeatures)
`outcome`	vector of outcomes (factor)
`nLayers`	number of hidden layers in network
`nUnits`	number of units in each hidden layer
`seed`	seed for intilialisng parameters