[R] by function ??

Matthew Dowle mdowle at mdowle.plus.com
Wed Dec 23 14:26:37 CET 2009


You asked how to 'create confidence intervals around the median'. Since 
intervals was plural and you asked within the context of 'by' LEAID then 
that made sense and I guessed you meant 'confidence interval around the 
median ratio of each group'. I already assumed you had posted a small subset 
of data, which is why I wrote 'with enough data' to explain the warning 
message.

Maybe someone else on the list understands you better and can help, but if 
not then please be much more specific. If you can't 'get it to run', then 
what exactly is 'it' and whats the error message? If there is no error then 
what is the result returned and what do you want instead?  Please follow the 
posting guidelines as you did at the start of this thread (including the R 
version details).

"L.A." <romsa at millect.com> wrote in message 
news:1261529371221-977494.post at n4.nabble.com...
>
> Thanks, but that produces what I think is an estimated interval.
>   I really want to use the above formula. I just can't figure out how to
> get it to run by the LEAID.
> It does require 9 observations to produce an interval, but I was showing a
> sample.
> Thanks again.
> L.A.
>
>
> Matthew Dowle-3 wrote:
>>
>> Maybe this (with enough data for a CI) ? :
>>
>>> Dataset = data.table(Dataset)
>>> Dataset[,as.list(wilcox.test(ratio,conf.int=TRUE)$conf.int),by="LEAID"]
>>      LEAID              V1      V2
>> [1,]  6307   0.7200000   0.92
>> [2,]  8300   0.5678462   0.83
>> Warning messages:
>> 1: In switch(alternative, two.sided = { :
>>   Requested conf.level not achievable
>>
>>
>> "L.A." <romsa at millect.com> wrote in message
>> news:1261418507079-976414.post at n4.nabble.com...
>>>
>>>
>>> Well, I'm back again.
>>>   Thanks for all the help. Besides working, it's helping me begin to
>>> understand how these functions work.
>>> I still have trouble reading or following the process of a function,
>>> which
>>> brings my next question:
>>>
>>> Dataset:
>>>
>>> LEAID     ratio
>>> 3 6307     0.7200000
>>> 1 6307     0.7623810
>>> 2 6307     0.8600000
>>> 4 6307     0.9200000
>>> 5 8300     0.5678462
>>> 7 8300     0.7700000
>>> 6 8300     0.8300000
>>>
>>> Here I want to create confidence intervals around the median. My hang-up
>>> now
>>> is how to do pbinom() by LEAID.
>>>
>>> Here's my script.
>>>
>>> Dataset$obs <- with(Dataset, ave(ratio, LEAID, 
>>> FUN=function(x)length(x)))
>>> obs <-by(Dataset[ , "ratio"], LEAID, length)
>>> obsn <-by(Dataset[ , "obs"], LEAID, FUN=function(x)1:max(x))
>>>
>>> q <- obsn
>>> n <- obs
>>> pbinom <-pbinom(q, n, .5)
>>> pbinom <- with(Dataset, by(obs, LEAID, FUN=function(x)pbinom(q, n, .5)))
>>> x <- pbinom
>>> x.nolag=x[2:n]
>>> x.lag1=x[1:(obs-1)]
>>> my.new=cbind(x.nolag,x.lag1)
>>> x.fill= cbind(0,0)
>>> my.dat.fr=rbind(x.fill, my.new)
>>> my.dat.fr
>>> pr=1 - (2 * my.dat.fr)
>>> cil=ifelse(pr >= .95, Dataset[,"ratio"], 0)
>>> cil
>>> max=sort(cil)
>>> LowerCI=round(max(max),digits=2)
>>> LowerCI
>>> cn=ifelse(pr >= .95, Dataset[,"ObsNumber"], 0)
>>> cmax=sort(cn)
>>> c=max(cmax)
>>> prb=ifelse(pr >= .95, pr, 999)
>>> con=sort(prb)
>>> Confidence=min(con)
>>> o= n -( c - 1)
>>> ch=ifelse (q <= o , Dataset[,"Ratio"], 0)
>>> UpperCI=round(max(ch),digits=2)
>>> UpperCI
>>>
>>>
>>> I hope your not growing tired of my questions, I try reading all I can
>>> find,
>>> but I often can't seem understand it until I can actually see it in
>>> process.
>>> Thanks,
>>> L.A.
>>> -- 
>>> View this message in context:
>>> http://n4.nabble.com/by-function-tp955789p976414.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.
>>
>>
>
> -- 
> View this message in context: 
> http://n4.nabble.com/by-function-tp955789p977494.html
> Sent from the R help mailing list archive at Nabble.com.
>




More information about the R-help mailing list