Learn R Programming

TeachNet (version 0.7.1)

Fits Neural Networks to Learn About Backpropagation

Description

Can fit neural networks with up to two hidden layer and two different error functions. Also able to handle a weight decay. But just able to compute one output neuron and very slow.

Copy Link

Version

Install

install.packages('TeachNet')

Monthly Downloads

28

Version

0.7.1

License

GPL (>= 2)

Maintainer

Georg Steinbuss

Last Published

November 27th, 2018

Functions in TeachNet (0.7.1)

computeOutput2

Computes output
TeachNet

Fits the neural network
createWeights1

Creates random weights
createWeights2

Creates random weights
crossEntropy

Cross entropy
logistic

Logistic function
confusion

Computes confusion matrix
is.decay

Checks for correct input
logistic.differential

Differential of logistic function
sumSquaredError

Sums up squared error
find.Threshold

Finds best threshold
transformPrediction

Transforms prediction
is.err

Checks for correct input
is.learn

Checks for correct input
is.numberOfNeurons

Checks for correct input
is.sample

Checks for correct input
Weights-class

Weights objects
Weights2-class

Weights2 objects
is.sampleLeng

Checks for correct input
fitTeachNet1

One step in backpropagation
squaredError

Computes squared error
fitTeachNet2

One step in backpropagation
sumCrossEntropy

Sums up cross entropy
is.stepMax

Checks for correct input
computeGrad2

Computes a gradient
computeOutput1

Computes output
is.thres.error

Checks for correct input
is.acct

Checks for correct input
is.data

Checks for correct input
predict.Weights

Computes prediction
predict.Weights2

Computes prediction
accuracy.me

Computes accuracy
computeGrad1

Computes a gradient
TeachNet-package

Fit neural networks with up to 2 hidden layers and one output neuron