[R] functions on rows or columns of two (or more) arrays

[Dennis Murphy]

Hi:

Here's one approach:

a=matrix(1:50,nrow=10)
a2=floor(jitter(a,amount=50))

# Write a function to combine the columns of interest
# into a data frame and fit a linear model
regfn <- function(k) {
     rdf <- data.frame(x = a[k, ], y = a2[k, ])
     lm(y ~ x, data = rdf)
   }

# Use lapply() to run regfn() recursively along
# the rows of a and a2:
modlist <- lapply(seq_len(nrow(a)), regfn)

# I prefer plyr for extraction of output from a list of models.
# Here are a few examples:

library('plyr')
# Extract the R^2 values
ldply(modlist, function(m) summary(m)$r.squared)
# Extract the residuals
laply(modlist, function(m) resid(m))
# Extract the estimated model coefficients
ldply(modlist, function(m) coef(m))
# Extract the coefficient summary tables as a list
llply(modlist, function(m) summary(m)$coefficients)

In the anonymous functions, the argument m refers to an arbitrary lm
object, so you can do to it what you would with any given lm object;
all you're doing is abstracting the process.

HTH,
Dennis

On Thu, Aug 4, 2011 at 2:17 PM, Jim Bouldin <bouldinjr at gmail.com> wrote:
> I realize this should be simple, but even after reading over the several
> help pages several times, I still cannot decide between the myriad "apply"
> functions to address it.  I simply want to apply a function to all the rows
> (or columns) of the same index from two (or more) identically sized arrays
> (or data frames).
>
> For example:
>
>> a=matrix(1:50,nrow=10)
>> a2=floor(jitter(a,amount=50))
>> a
>      [,1] [,2] [,3] [,4] [,5]
>  [1,]    1   11   21   31   41
>  [2,]    2   12   22   32   42
>  [3,]    3   13   23   33   43
>  [4,]    4   14   24   34   44
>  [5,]    5   15   25   35   45
>  [6,]    6   16   26   36   46
>  [7,]    7   17   27   37   47
>  [8,]    8   18   28   38   48
>  [9,]    9   19   29   39   49
> [10,]   10   20   30   40   50
>> a2
>      [,1] [,2] [,3] [,4] [,5]
>  [1,]   31 56 -29 -13 10
>  [2,]   38   61   71   55    9
>  [3,]  -29   38   47   12   38
>  [4,]   12    2   43   39   93
>  [5,]  -43   23  -23   62    1
>  [6,]  -13   61   55   11    2
>  [7,]  -42    1   38   12    8
>  [8,]  -13   -6  -18   16   95
>  [9,]  -19   -2   78   33    1
> [10,]   20 -16 -11 19 17
>
> if I try the following for example:
> apply(a,1,function(x) lm(a~a2))
>
> I get 10 identical repeats (except for the list indexer) of the following:
>
> [[1]]
>
> Call:
> lm(formula = a ~ a2)
>
> Coefficients:
>             [,1]       [,2]       [,3]       [,4]       [,5]
> (Intercept)   8.372135  18.372135  28.372135  38.372135  48.372135
> a21          -0.006163  -0.006163  -0.006163  -0.006163  -0.006163
> a22          -0.093390  -0.093390  -0.093390  -0.093390  -0.093390
> a23           0.009315   0.009315   0.009315   0.009315   0.009315
> a24          -0.015143  -0.015143  -0.015143  -0.015143  -0.015143
> a25          -0.026761  -0.026761  -0.026761  -0.026761  -0.026761
>
> ...Which is clearly very wrong, in a number of ways.  If I try by columns:
> apply(a,2,function(x) lm(a~a2))
> ...I get exactly the same result.
>
> So, which is the appropriate apply-type function when two arrays (or
> d.f.'s?) are involved like this? Or none of them and some other approach
> (other than looping which I can do but which I assume is not optimal)?
> Thanks for any help.
> --
> Jim Bouldin, PhD
> Research Ecologist
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
>