Actually, I think it's pretty trivial if you do it in a smarter way
than I previously suggested. I found this by reading ?levels (RTFM,
Bert!)
> z <- factor(letters[1:3])
> levels(z)[1:2]<- "d" ## no hardcoding names; just use indices
> z
[1] d d c
Levels: d c
Cheers,
Bert
On Fri, Aug 9, 2013 at
Thanks Bert. I guess I was just wondering if there was a way to create
the new factors automatically without me having to hard code the level
names manually in my R code.
Rgds
Claus
On Fri, Aug 9, 2013 at 3:42 PM, Bert Gunter wrote:
> ... or if you want to keep the unchanged levels the same:
>
... or if you want to keep the unchanged levels the same:
zz <- factor(ifelse( z %in% c("a", "b"),"d" ,levels(z)[z]))
-- Bert
On Fri, Aug 9, 2013 at 7:35 AM, Bert Gunter wrote:
> If I understand what you mean, just recode them.
>
> z <- factor(letters[1:3])
> z
> zz <- factor(ifelse( z %in% c("
If I understand what you mean, just recode them.
z <- factor(letters[1:3])
z
zz <- factor(ifelse( z %in% c("a", "b"),"d" ,z))
zz
Cheers,
Bert
On Fri, Aug 9, 2013 at 7:10 AM, Claus O'Rourke wrote:
> Hello R-Help,
> I have a variable with > 32 levels and I'd like to split this into two
> variable
Hello R-Help,
I have a variable with > 32 levels and I'd like to split this into two
variables such that both new variables have >= 32 variables. This is
to handle the limit of 32 level predictor variables in R's Random
Forest implementation. Might someone be able to suggest an elegant way
to do th
5 matches
Mail list logo
