> ---------- Forwarded message ----------
> Date: Mon, 18 Jul 2011 10:17:00 -0700 (PDT)
> From: KHOFF <kuphoff at gmail.com>
> To: r-help at r-project.org
> Subject: [R] cforest - keep.forest = false option?
>
> Hi,
>
> I'm very new to R. I am most interested in the variable importance
> measures
> that result from randomForest, but many of my predictors are
> highly
> correlated. My first question is:
>
> 1. do highly correlated variables render variable importance
> measures in
> randomForest invalid?
>
that depends on your idea of "valid". A number of papers
published over the last years explore this question and
you should read the relevant literature first.
> and 2. I know that cforest is robust to highly correlated
> variables,
> however, I do not have enough space on my machine to run cforest.
> I used the
> keep.forest = false option when using randomForest and that solved
> my space
> issue. Is there a similar option for cforest (besides
> savesplitstats > FALSE, which isn't helping)
no. party was designed as a flexible research tool and is
not optimized wrt speed or memory consumption.
Best,
Torsten
>
> below is my code and error message
>
> Thanks in advance!
>
>> fit <- cforest(formula = y ~ x1 + x2+ x3+ x4+ x5+
> + x6+ x7+ x8+ x9+ x10, data=data, control=
> cforest_unbiased(savesplitstats > FALSE, ntree = 50, mtry = 5)
>
> 1: In mf$data <- data :
> Reached total allocation of 3955Mb: see help(memory.size)
> 2: In mf$data <- data :
> Reached total allocation of 3955Mb: see help(memory.size)
>
>
> --
> View this message in context:
>
http://r.789695.n4.nabble.com/cforest-keep-forest-false-option-tp3675921p3675921.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
