transWeightListToVect: Transform the List of Optimal Weights to a Numeric Vector

In nnetpredint: Prediction Intervals of Multi-Layer Neural Networks

Description

This function transforms the list of weight parameters,typically the output of neuralnet package (nn$weights) to a single numeric vector in specific order, which will be used as the input to the prediciton interval function. The order of weight parameters is the same as the output of nnet package (nnet$wts).

Usage

transWeightListToVect(wtsList,m = 2)

Arguments

wtsList	the list of weights found by neuralnet package (nn$weights). For nnet weights output(nnet$wts), there is no need for transformation and it can be used directly by the prediction interval method.
m	the number of layers of the neural networks, which is the number of hidden layer plus + 1. m is default to 2 for single hidden layer networks.

Value

the numeric vector of optimal weights found by neural networks, dimension 1 * nPara.

Author(s)

Xichen Ding <[e-mail]>

References

neuralnet package by Stefan Fritsch, Frauke Guenther

See Also

nnetPredInt

nnetpredint documentation built on May 2, 2019, 6:12 p.m.