tournament: Choose parents using tournament selection. Each time,...

In kunaljaydesai/GA2: Implement Genetic Variable Selection Algorithm

Description

Choose parents using tournament selection. Each time, randomly pick k models from the previous generation and the best-performed model will be selected as the parent. Repeat this process until all slots for parents are filled.

Usage

tournament(models, k)

Arguments

models	dataframe with models (specified by the index of selected variables) and their respective fitness value. Object returned by ranked_models().
k	size of each round of selection. Must be an integer smaller than generation size.

Value

The selected parents(list of ceiling(nrow(models)/2) lists, each list contains two parents)

Examples

x <- mtcars[-1]
y <- unlist(mtcars[1])
initial_index <- initialize_parents(10,20)$index
rankedmodels <- ranked_models(initial_index,x,y)
tournament(rankedmodels,k = 2)

kunaljaydesai/GA2 documentation built on May 30, 2019, 2:15 p.m.