R caret random forest script failing to load

Asked Apr 28 '18 at 16:07

Active Apr 28 '18 at 16:07

Viewed 43 times

I'm trying to build a machine learning model using caret's build in rf method.

when i do my cross-validation it uses about 6 GB of Ram but nothing happens?

My data set is 11.000 obs. with 16 variables. all of them are either intergers, numeric, factor. My code is:

trainControl <- trainControl(method ="cv", number = 10)
metricType <- "Accuracy"

## Random Forest ##

set.seed(10)
model.rf <- train(Udmeldt~., data = TrainData, method="rf", metric = metricType, trControl = trainControl)

i'm running it on a win 10, with 8 GB ram.

asked Apr 28 '18 at 16:07

William Parker

Our dataset has a column with names for the 11.000 obs. these are all factor variables does that play a role? – William Parker Apr 28 '18 at 16:18
What do you mean by "nothing happens"? – Hong Ooi Apr 28 '18 at 16:19
it just keeps loading without returning a output or error message – William Parker Apr 28 '18 at 16:57
1

You're probably running out of memory and thrashing the hard disk. Try tweaking the RF parameters to reduce the model size, eg limit the no. of trees to 20 or 50. – Hong Ooi Apr 28 '18 at 17:00
How many variables are factors and what is the total amount of levels in those factors? Try using `method = "ranger"` – missuse Apr 29 '18 at 08:13
This [question](https://stackoverflow.com/questions/6449588/r-memory-management-advice-caret-model-matrices-data-frames) is relevant. Be sure to check out the comments too. – missuse Apr 29 '18 at 08:23
Thanks Hong Ooi for the help. i limmited the amount of trees and it helped! – William Parker Apr 29 '18 at 13:42

R caret random forest script failing to load

0 Answers0