Skip to contents

Run the BioGA genetic algorithm

Source: R/bioga.R
bioga.Rd

User-facing wrapper around bioga_main_cpp() with sensible defaults and input validation. The algorithm minimises two objectives; in "binary" mode pass class labels so that the first objective measures the predictive signal of the selected genes.

Usage

bioga(
  genomic_data,
  population_size = 50,
  num_generations = 20,
  crossover_rate = 0.9,
  mutation_rate = 0.1,
  num_parents = 20,
  num_offspring = 20,
  num_to_replace = 10,
  weights = c(1, 0.5),
  labels = NULL,
  clusters = NULL,
  network = NULL,
  seed = NULL,
  eta_c = 20,
  representation = c("real", "binary")
)

Arguments

genomic_data

Numeric matrix of genomic data (rows: genes, columns: samples).

population_size

Number of individuals in the population.

num_generations

Number of generations to run.

crossover_rate

Probability of crossover.

mutation_rate

Base probability of mutation.

num_parents

Number of parents selected per generation.

num_offspring

Number of offspring generated per generation.

num_to_replace

Maximum number of offspring admitted per generation.

weights

Numeric vector of objective weights (length 1 or 2).

labels

Optional class labels, one per sample (integer, numeric, factor, character or logical). Required when representation = "binary".

clusters

Optional vector of gene cluster assignments.

network

Optional symmetric gene network adjacency matrix.

seed

Optional integer random seed for reproducibility.

eta_c

SBX distribution index (default: 20).

representation

Either "real" or "binary".

Value

A list with population, fitness, history, best, best_fitness, pareto_population and pareto_fitness.

See also

bioga_fitness(), bioga_main_cpp()

Examples

genomic_data <- matrix(rnorm(200), nrow = 20, ncol = 10)
result <- bioga(genomic_data, population_size = 20, num_generations = 5)