0

I have a question about the binning of continuous variables by cut2 from Hmisc in R and why in some cases the cutpoints are not treated the same.

I am aware of this question, especially the answer from Christopher Bottoms, but it doesnt address my question as to why cut2 doesnt seem to respect the cutpoints supplied to it in certain circumstances.

Given

v<-seq(1:12)
v

I want to supply a list of cutpoints (a,b,c..,y,z) and have a numeric variable binned into the ranges [-Inf,b), [b,c),...[y,Inf]

This seems to work fine.

    cuts<-cut2(v,g = 4,onlycuts = TRUE)
    cuts[1]<- -Inf
    cuts[length(cuts)]<- Inf 
    cuts
    > cuts
    [1] -Inf    4    7   10  Inf

table(cut2(v,cuts = cuts))
> table(cut2(v,cuts = cuts))

[-Inf,   4) [   4,   7) [   7,  10) [  10, Inf] 
          3           3           3           3

But not this. How to accomplish a binning based on a user defined rule like this?

cuts<-cut2(v,g = 7,onlycuts = TRUE)
cuts[1]<- -Inf
cuts[length(cuts)]<- Inf 
cuts

> cuts
[1] -Inf    3    5    7    8   10  Inf

table(cut2(v,cuts = cuts))

> table(cut2(v,cuts = cuts))

[-Inf,   3) [   3,   5) [   5,   7)           7 [   8,  10) [  10, Inf] 
          2           2           2           1           2           3
Community
  • 1
  • 1
B_Miner
  • 1,840
  • 4
  • 31
  • 66
  • 2
    What exactly is the question here? I'm not sure I see a problem with the output. Are you just confused by the lone 7? Try `table(cut2(v,cuts = cuts, oneval=FALSE))` – MrFlick Nov 27 '15 at 20:54
  • I think that was what I was missing! I will need to test some more but seemed to work! – B_Miner Nov 27 '15 at 23:26

0 Answers0