R help - Apr 2010 - Binning Question

Hi,

I'm trying to setup some complicated binning with statistics and could
use a little help.

I've found the bin2 function from the ash package, but it doesn't do
everything I need.  My intention is to copy some of their code and then
modify as needed.

I have a vector of two columns:

head(data)
              r1          r2
[1,]  0.03516559  0.03102128
[2,]  0.02162539  0.14847034
[3,]  0.02210339  0.06539623
[4,] -0.07547792 -0.08859678
[5,]  0.03655620  0.05412436
[6,]  0.06513828  0.06053050


I'd like to create a 2 dimension list of bins with the frequency counts
for each bin.  The bin2 function does this.  Then it gets interesting.

I'd like to add a column to my vector that has the "bin label" for
the
bin that row would belong to.  (I can see how to do this with lots of
nasty loops and greater-than, less-than calculations, but that gets
messy.)  There must be an easier way.

So, If I made 10 bins for each column (r1,r2), I'd have 100 bins. 
(bin1, bin2, bin3, etc.)  I want to label each ROW in my data set with
the bin it would belong to.  (I intend to do more work with them after
this, but this starts.  Each row gets transformed depending on the bin
it belongs to, etc..)

Thanks,

-N

On Apr 12, 2010, at 9:07 PM, Noah Silverman wrote:
> Hi,
>
> I'm trying to setup some complicated binning with statistics and could
> use a little help.
>
> I've found the bin2 function from the ash package, but it doesn't
do
> everything I need.  My intention is to copy some of their code and  
> then
> modify as needed.
>
> I have a vector of two columns:
>
> head(data)
>              r1          r2
> [1,]  0.03516559  0.03102128
> [2,]  0.02162539  0.14847034
> [3,]  0.02210339  0.06539623
> [4,] -0.07547792 -0.08859678
> [5,]  0.03655620  0.05412436
> [6,]  0.06513828  0.06053050
>
>
> I'd like to create a 2 dimension list of bins with the frequency  
> counts
> for each bin.  The bin2 function does this.  Then it gets interesting.
>
> I'd like to add a column to my vector that has the "bin
label" for the
> bin that row would belong to.  (I can see how to do this with lots of
> nasty loops and greater-than, less-than calculations, but that gets
> messy.)  There must be an easier way.
Lets say you used the example in bin2:

dat <- as.data.frame(matrix( rnorm(200), 100 , 2)) # bivariate normal  
n=100

ab <- matrix( c(-5,-5,5,5), 2, 2) # interval [-5,5) x [-5,5)
nbin <- c( 20, 20) # 400 bins
bins <- bin2(dat, ab, nbin) # bin counts,ab,nskip

dat$r1.cat <- cut(dat[,1],  breaks=seq(ab[1,1], ab[1,2],  
length.out=nbin[1]+1 ) )
dat$r2.cat <- cut(dat[,2], breaks=seq(ab[1,1], ab[1,2],  
length.out=nbin[1]+1))
dat$bicat <- with(dat, paste( as.numeric(r1.cat), as.numeric(r2.cat),  
sep="."))

Or leave off the as.numeric if you want the labels to be more "cut"- 
like.

>
> So, If I made 10 bins for each column (r1,r2), I'd have 100 bins.
> (bin1, bin2, bin3, etc.)  I want to label each ROW in my data set with
> the bin it would belong to.  (I intend to do more work with them after
> this, but this starts.  Each row gets transformed depending on the bin
> it belongs to, etc..)
>
> Thanks,
>
> -N
>
> ______________________________________________
> 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.