In gmcmacran/NeuralNetworkSimulatoR: Creates Ideal Data for a Neural Network

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Introduction

In this vignette, data for a deep feed forward network is created. Please read the README file before reading this vignette.

Matrices and Activations

First a list of matrices must be made. The elements of each matrix are the Betas of the network. The dimensions are associated with the network's structure. Here is an example of creating ideal data for a network with 3 hidden layers. The first hidden layer has 10 nodes. The second hidden layer has 8 nodes. The third hidden layer has 5 nodes.

library(NeuralNetworkSimulatoR)

set.seed(1)
M <- list(
  matrix(rnorm(12*10, mean = 0, sd = 1), nrow = 12, ncol = 10),
  matrix(rnorm(10*8, mean = 0, sd = 1), nrow = 10, ncol = 8),
  matrix(rnorm(8*5, mean = 0, sd = 1), nrow = 8, ncol = 5),
  matrix(rnorm(5*1, mean = 0, sd = 1), nrow = 5, ncol = 1)
  )

nrow in the first matrix (12 in the above) is the number of columns in the data set (see below). ncol of the first matrix must be nrow of the second matrix (10 above). This repeats until the last matrix which must have only one column.

Next activation functions for each hidden layer are selected. In this example, the first three activations are relu and the last is linear.

A <- list(relu_R, 
          relu_R,
          relu_R,
          linear_R)

Creating Data

With the matrices and activations created, a data set containing 12 input variables following a normal distribution is made.

simData <- simulate_regression_data(
  rows = 1000L,
  N = 12L, U = 0L, C = 0L,
  matrices = M, activations = A,
  noise = 0
)

head(simData)

A data set with only categorical variables can also be created.

simData <- simulate_regression_data(
  rows = 1000L,
  N = 0L, U = 0L, C = 12L,
  matrices = M, activations = A,
  noise = 0
)

head(simData)

Or a mixture of normal, uniform and categorical variables.

simData <- simulate_regression_data(
  rows = 1000L,
  N = 6L, U = 3L, C = 3L,
  matrices = M, activations = A,
  noise = 0
)

head(simData)

gmcmacran/NeuralNetworkSimulatoR documentation built on March 23, 2020, 12:51 a.m.