tournament_selection: tournament_selection

In Skjemaa/GA: Selects covariates for use in regression using genetic algorithm

Description

A parents selection mechanism. The tournament selection splits the parents population in k non overlapping subgroups and takes the fittest individual in each group then takes n-k random subgroups (n the number of parents) and takes the fittest individual in each of these random subgroups

Usage

tournament_selection(y, dataset, individuals, k = 4, n_var,
  objective = "AIC")

Arguments

y	response variable
dataset	the dataframe containing the data
k	the number of subsequences that should not overlap
n_var	number of parents
objective	the objective function

Value

a list of n_var parents that are the fittest ones in each subgroup

Skjemaa/GA documentation built on May 3, 2019, 6:42 p.m.