Genetic algorithm using feature selection

Question

I am trying to use Caret Feature Selection using Genetic Algorithms, am getting an error message. My code is shown below :

set.seed(10)
trainIndex <- createDataPartition(iris$Species, p = .5, list = FALSE, times = 1)
trainData <- iris[trainIndex,-c(1,2)]
testData <- iris[-trainIndex,-c(1,2)]
trainX <-trainData[,-1]
testX <- testData[,-1]
y=trainData$Class
data(iris)
dim(iris) 
# [1] 150   5
head(iris,2)   
#   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
# 1          5.1         3.5          1.4         0.2  setosa 
# 2          4.9         3.0          1.4         0.2  setosa
registerDoParallel(4)
getDoParWorkers() [1] 4
utils:::menuInstallLocal()
# le package ‘GA’ a été décompressé et les sommes MD5 ont été vérifiées avec succés
ga_ctrl <- gafsControl(functions = rfGA, method = "cv", genParallel=TRUE, allowParallel = TRUE)
set.seed(10)
lev <- c("PS","WS")
system.time(rf_ga3 <- gafs(x = trainX, y = y, iters = 100, popSize = 20, levels = lev, gafsControl = ga_ctrl))
# Erreur dans gafs.default(x = trainX, y = y, iters = 100, popSize = 20, levels = lev,  :    
#    there should be the same number of samples in x and y Timing stopped at: 0 0 0

It looks like your y vector have the wrong size. Maybe you're sampling X and forgot to also sample Y. — Fernando
– Fernando, Commented Mar 17, 2017 at 14:09
Dear Rania, please show some effort in formatting your question. To address your question: @Fernando is right, that your are using complete y where as x only uses indices for training. Therein lies the mistake. — Drey
– Drey, Commented Mar 17, 2017 at 14:10

Fernando · Accepted Answer · 2017-03-17 14:13:42Z

1

It's a common mistake. You sampled X but forgot to sample Y. Do something like:

ytrain = y[trainIndex]
ytest = y[-trainIndex]

answered Mar 17, 2017 at 14:13

Fernando

7,9256 gold badges53 silver badges85 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Genetic algorithm using feature selection

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related