Hi Everyone,
Looks like I found the solution, the mode function is now changed to:
mode <- function(x){
cell.mode<-as.numeric(names(which.max(table(x))))
if(length(cell.mode)==0){
cell.mode <- NA
}
cell.mode
}
Then you can take out the useNA = 'ifAny' in the apply line. Thanks everyone
and especially Henrique for all your help!
Kind regards,
Greg
Begin forwarded message:
> From: Gregory Ryslik <rsa...@comcast.net>
> Date: September 21, 2010 4:08:33 PM EDT
> To: Henrique Dallazuanna <www...@gmail.com>
> Cc: r-h...@stat.math.ethz.ch
> Subject: Re: [R] mode across lists of matrices
>
> Hey,
>
> Looks like I spoke to soon last time. It's almost there but not quite. Now If
> there are more NA's than any other number, it returns NA. That's not the
> correct behavior though as I want to return NA only if all the numbers are
> NA. For example, in the situation below, I want it to return 5 for element
> [1,1]. Thanks though for all your help!
>
> If I take out the "useNA = ifany" then I get a "numeric, 0" if all were NA
> and the correct number if at least some were not NA. Is there a way I can
> test for the "numeric, 0" at the end when the matrix is constructed and swap
> them out? Or is there a simpler way?
>
> Thanks again,
> Greg
>
> mymats <- vector('list', 4)
> set.seed(246)
>
> # Generate a list of five 3 x 3 matrices
> for(i in 1:4) mymats[[i]] <- matrix(sample(1:9), nrow = 3)
>
> mymats[[1]][1,1]<-NA
> mymats[[1]][1,2]<-3
> mymats[[1]][1,3]<-1
> mymats[[2]][2,1]<-2
> mymats[[1]][2,2]<-3
> mymats[[1]][2,3]<-1
> mymats[[1]][3,1]<-2
> mymats[[1]][3,2]<-2
> mymats[[1]][3,3]<-3
>
> mymats[[2]][1,1]<-NA
> mymats[[2]][1,2]<-3
> mymats[[2]][1,3]<-2
> mymats[[2]][2,1]<-1
> mymats[[2]][2,2]<-2
> mymats[[2]][2,3]<-2
> mymats[[2]][3,1]<-1
> mymats[[2]][3,2]<-3
> mymats[[2]][3,3]<-2
>
> mymats[[3]][1,1]<-NA
> mymats[[3]][1,2]<-2
> mymats[[3]][1,3]<-2
> mymats[[3]][2,1]<-2
> mymats[[3]][2,2]<-3
> mymats[[3]][2,3]<-1
> mymats[[3]][3,1]<-2
> mymats[[3]][3,2]<-2
> mymats[[3]][3,3]<-3
>
> mymats[[4]][1,1]<-5
> mymats[[4]][1,2]<-1
> mymats[[4]][1,3]<-NA
> mymats[[4]][2,1]<-3
> mymats[[4]][2,2]<-2
> mymats[[4]][2,3]<-3
> mymats[[4]][3,1]<-1
> mymats[[4]][3,2]<-2
> mymats[[4]][3,3]<-1
>
> mymats
>
>
> mode <- function(x){
> as.numeric(names(which.max(table(x,useNA = 'ifany'))))
> }
> answer <- apply(array(unlist(mymats), dim = c(dim(mymats[[1]]),
> length(mymats))), 1:2, mode)
>
>
>
> On Sep 21, 2010, at 3:27 PM, Henrique Dallazuanna wrote:
>
>> Use this function:
>>
>> mode <- function(x, ...)
>> as.numeric(names(which.max(table(x, useNA = 'ifany'))))
>>
>> On Tue, Sep 21, 2010 at 4:18 PM, Gregory Ryslik <rsa...@comcast.net> wrote:
>> Hi,
>>
>> Much better! That seems to work great. The only time it doesn't work is that
>> if all the elements are NA for a specific element. Then it gives me a
>> numeric(0) for that cell rather than NA.
>>
>> On Sep 21, 2010, at 2:30 PM, Henrique Dallazuanna wrote:
>>
>>> Try this:
>>>
>>> apply(array(unlist(mymats), dim = c(dim(mymats[[1]]), length(mymats))),
>>> 1:2, mode)
>>>
>>> The error was in c(length(mymats), dim(mymats[[1]]))
>>>
>>> On Tue, Sep 21, 2010 at 3:19 PM, Gregory Ryslik <rsa...@comcast.net> wrote:
>>> Ack, apologies for the previous email. What I meant to say is that the
>>> first element is calculated incorrectly (1,1) should be 2, instead it is 3.
>>> I've been staring at the code for two long. I've copied it in again for
>>> convenience.
>>>
>>> mymats <- vector('list', 4)
>>> set.seed(246)
>>>
>>> # Generate a list of 4 3 x 3 matrices
>>> for(i in 1:4) mymats[[i]] <- matrix(sample(1:9), nrow = 3)
>>>
>>> mymats[[1]][1,1]<-3
>>> mymats[[1]][1,2]<-3
>>> mymats[[1]][1,3]<-1
>>> mymats[[2]][2,1]<-2
>>> mymats[[1]][2,2]<-3
>>> mymats[[1]][2,3]<-1
>>> mymats[[1]][3,1]<-2
>>> mymats[[1]][3,2]<-2
>>> mymats[[1]][3,3]<-3
>>>
>>> mymats[[2]][1,1]<-2
>>> mymats[[2]][1,2]<-3
>>> mymats[[2]][1,3]<-2
>>> mymats[[2]][2,1]<-1
>>> mymats[[2]][2,2]<-2
>>> mymats[[2]][2,3]<-2
>>> mymats[[2]][3,1]<-1
>>> mymats[[2]][3,2]<-3
>>> mymats[[2]][3,3]<-2
>>>
>>> mymats[[3]][1,1]<-NA
>>> mymats[[3]][1,2]<-2
>>> mymats[[3]][1,3]<-2
>>> mymats[[3]][2,1]<-2
>>> mymats[[3]][2,2]<-3
>>> mymats[[3]][2,3]<-1
>>> mymats[[3]][3,1]<-2
>>> mymats[[3]][3,2]<-2
>>> mymats[[3]][3,3]<-3
>>>
>>> mymats[[4]][1,1]<-2
>>> mymats[[4]][1,2]<-1
>>> mymats[[4]][1,3]<-2
>>> mymats[[4]][2,1]<-3
>>> mymats[[4]][2,2]<-2
>>> mymats[[4]][2,3]<-3
>>> mymats[[4]][3,1]<-1
>>> mymats[[4]][3,2]<-2
>>> mymats[[4]][3,3]<-1
>>>
>>> mymats
>>>
>>>
>>> mode <- function(x){
>>> as.numeric(names(which.max(table(x))))
>>> }
>>> apply(array(unlist(mymats), dim = c(length(mymats), dim(mymats[[1]]))),
>>> 1:2, mode)
>>>
>>>
>>> On Sep 21, 2010, at 10:08 AM, Henrique Dallazuanna wrote:
>>>
>>>> Try this:
>>>>
>>>> mode <- function(x, ...)
>>>> as.numeric(names(which.max(table(x))))
>>>> apply(array(unlist(mymats), dim = c(length(mymats), dim(mymats[[1]]))),
>>>> 1:2, mode)
>>>>
>>>>
>>>> On Tue, Sep 21, 2010 at 10:47 AM, Gregory Ryslik <rsa...@comcast.net>
>>>> wrote:
>>>> Hi Everyone,
>>>>
>>>> I am interested in taking the mode over several thousand matrices. I show
>>>> an example below. For the [1,1] entry of my "mode" matrix that I want to
>>>> create I would like to have a "2". For the [1,2] entry I would want a 2.
>>>> For the [2,2] entry it would be 4 and so forth. Earlier, I was working
>>>> with continuous cases and thus each (n,m) element was simply an average. I
>>>> was able to then do element-wise addition and counting using the "Reduce"
>>>> function and then the average would be totalsum/totalcount where the NA
>>>> terms were discounted. Here, it's not exactly a binary case so Reduce
>>>> doesn't quite work as well. I'm open to suggestions but would as always
>>>> like to avoid long loops as that will significantly bump up running time
>>>> over several thousand trees. Similarly, I would not like to do a lot of
>>>> sorts to find the mode either...
>>>>
>>>> Thanks for your help!
>>>>
>>>> mymats
>>>> [[1]]
>>>> [,1] [,2] [,3]
>>>> [1,] 0 2 1
>>>> [2,] 2 3 3
>>>> [3,] 2 1 2
>>>>
>>>> [[2]]
>>>> [,1] [,2] [,3]
>>>> [1,] 1 2 4
>>>> [2,] 2 4 4
>>>> [3,] 3 4 5
>>>>
>>>> [[3]]
>>>> [,1] [,2] [,3]
>>>> [1,] 2 3 1
>>>> [2,] 3 4 2
>>>> [3,] 5 1 3
>>>>
>>>> [[4]]
>>>> [,1] [,2] [,3]
>>>> [1,] 2 4 2
>>>> [2,] 1 NA 2
>>>> [3,] 2 3 1
>>>>
>>>> [[5]]
>>>> [,1] [,2] [,3]
>>>> [1,] NA 2 1
>>>> [2,] 2 4 1
>>>> [3,] 1 3 2
>>>> ______________________________________________
>>>> R-help@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.
>>>>
>>>>
>>>>
>>>> --
>>>> Henrique Dallazuanna
>>>> Curitiba-Paraná-Brasil
>>>> 25° 25' 40" S 49° 16' 22" O
>>>
>>>
>>>
>>>
>>> --
>>> Henrique Dallazuanna
>>> Curitiba-Paraná-Brasil
>>> 25° 25' 40" S 49° 16' 22" O
>>
>>
>>
>>
>> --
>> Henrique Dallazuanna
>> Curitiba-Paraná-Brasil
>> 25° 25' 40" S 49° 16' 22" O
>
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
