R/mutation.R

In lixiao0214/GA: Genetic Algorithms package for STAT243 course

generate_mutation <- function(input,
                              mutation_rate,
                              main_dataset,
                              regression_target
                              ){
  #' Generate offsprings after mutation
  #'
  #' @description generate_mutation is used to generate offsprings after mutations based on particular mutation rate.
  #' And it will give the values of goodness of fit for each offspring.
  #'
  #' @author Xiao Li
  #'
  #' @usage generate_mutation(input, mutation_rate, main_dataset)
  #'
  #' @param input A matrix contains all the populations which are required to mutate.
  #' The number in each cell should be either 0 or 1. Each row represents an individual.
  #' Each column represents the variable.
  #' @param mutation_rate The chance for each variable in each individual to change the value.
  #' @param main_dataset The dataset used to run the regression model.
  #' And its model help to get the corresponding value of goodness of fit.
  #' @param regression_target character: column of df regressed on other covariates in df
  #'
  #' @examples # Generate random input and choose a specific dataset in R
  #' regression_target <- "V1"
  #' toy_datasets <- generate_toy_dataset(regression_target = regression_target)
  #' main_dataset <- toy_datasets$main_dataset
  #' population <- as.matrix(toy_datasets$population)
  #' generate_mutation(population, 0.01, main_dataset,regression_target)

  stopifnot(is.matrix(input))

  if (mutation_rate < 0 || mutation_rate > 1) {
    stop("mutation_rate should be a number between 0 and 1.")
  }

  # Generate a copy of the input without regression_target
  after_mutation <- input

  # Perform mutation
  # Generate random number from Unif(0,1) for each cell in the input
  # It the random number is smaller than mutation rate
  # The corresponding cell will mutate
  rand <- runif(length(after_mutation))
  index <- which(rand < mutation_rate)
  after_mutation[index] <- 1 - after_mutation[index]

  # Get goodness of fit for each offspring
  offsprings <- compute_population_goodness_of_fit(data = main_dataset,
                                                   population = after_mutation,
                                                   regression_target = regression_target,
                                                   )
  return(offsprings)
}