get_best_nn: Calculates the best neural network architecture that fixes a...

In cvazquezlos/NNEvol: A Simple Method to Obtain the Best Neural Network that Fixes a given Problem

Description

Calculates the best neural network architecture that fixes a given problem

Usage

get_best_nn(data, input, output, train = 0.7, validation = 0.2,
  test = 0.1, train_mode = 0, train_final_nn = FALSE, seed = 123,
  population_size = 30, generation_number = 30)

Arguments

data	entry dataset
input	number of inputs within the dataset
output	number of outputs within the dataset
train	proportion of data that is used to train each individual
validation	proportion of data that is used to validate each individual
test	proportion of data that is used to evaluate the final individual if option train_final_nn is TRUE
train_mode	mode of the training. 0: multi-class, 1: regression
train_final_nn	if the final individual is evaluated or not
seed	seed
population_size	size of the population
generation_number	number of generations

Value

List with both architecture and final loss

Examples

{
  data <- data.frame(X1 = c(1:10), X2 = c(11:20), X3 = c(1:10), Y1 = c(0:9),
                     stringsAsFactors = FALSE)
  get_best_nn(data, 3, 1, train_mode = 1, population_size = 5,
              generation_number = 2)
}

cvazquezlos/NNEvol documentation built on Nov. 4, 2019, 9:20 a.m.