doughnutRand

<p>V1 to V8 are the extra random features (in the range 0-1)
V9 and V10 are the original input features, normalized to 0-1
V11 is the output (TRUE / FALSE)</p>

datasets

A supervised transformation of datasets is performed.
The aim is similar to that of Principal Component Analysis (PCA), that is, to carry out data transformation and dimensionality reduction, but in a supervised way.
This is achieved by first training a 3-layer Multi-Layer Perceptron and then using the activations of the hidden layer as a transformation of the input features.
In fact, it takes advantage of the change of representation provided by the hidden layer of a neural network.
This can be useful as data pre-processing for Machine Learning methods in general, specially for those that do not work well with many irrelevant or redundant features.
It uses the nnet package under the hood.
Valls, J.M., Aler, R., Galvan, I.M., and Camacho, D. (2021). "Supervised data transformation and dimensionality reduction with a 3-layer multi-layer perceptron for classification problems". <doi:10.1007/s12652-020-02841-y>
Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986) "Learning representations by back-propagating errors" <doi:10.1038/323533a0>.

Ricardo Aler

nntrf

Supervised Data Transformation by Means of Neural Network Hidden
Layer

Jose Valls

Ines Galvan

David Camacho

doughnutRand function

<p>An object of class <code>data.frame</code> with 10000 rows and 11 columns.</p>

Format

Doughnut shaped dataset with 8 extra random attributes — doughnutRand

Doughnut shaped dataset with 8 extra random attributes

doughnutRand: Doughnut shaped dataset with 8 extra random attributes

Description

Usage

Arguments

Format

Examples