[R] apply regression to an array

[William Dunlap]

Since the model matrix, cbind(1,time) is the same for all your
response variables,
you can calculate this on one call to lm, but you have to rearrange the response
values so that each x,y set is in one column.  I think the following
function does it:

f <- function (time, y)
{
    stopifnot(length(dim(y)) == 3, dim(y)[3] == length(time))
    yMatrix <- matrix(aperm(y, c(3, 1, 2)), dim(y)[3])
    fit <- lm(yMatrix ~ time)
    aperm(array(fitted.values(fit), dim(y)[c(3, 1, 2)]), c(2,
        3, 1))
}

E.g.,
> fitted.values(lm(test1[2,5,]~time))
 1  2  3  4  5  6  7  8  9 10
 1  2  3  4  5  6  7  8  9 10
> f(time, test1)[2,5,]
 [1]  1  2  3  4  5  6  7  8  9 10


Bill Dunlap
TIBCO Software
wdunlap tibco.com


On Tue, Oct 6, 2015 at 10:42 AM, Adrienne Wootten <amwootte at ncsu.edu> wrote:
> R-Helpers,
>
> I've seen some similar threads about this question online, but not quite
> what I'm looking for.  I apologize in advance if someone's already answered
> this and I just can't find it online.
>
> Say that I have an array like test3 in the little example code I have below:
>
> test1 = array(rep(1:10,each = 25),dim=c(5,5,10))
> test2 = array(rnorm(250,0,0.35),dim=c(5,5,10))
> test3 = test1+test2 # array with 5 rows, 5 columns, 10 slices
>
> time=1:10
>
> Where the dimensions are x, y, and time.  What I'd like to do is run a
> regression (for the sake of this example, say lm) on each x,y in time.  So
> for a single cell the formula might be test3[1,1,]~time, but I'd like to
> that for all cells.  The only way I can immediately think of is to use a
> loop, but I'm wondering if there's a way to do this without a loop.
> Perhaps with tapply?
>
> I'm actually doing a fourth order regression with a much larger array, but
> this simple example illustrates the question I have.
>
> Many thanks for the help! Sorry if someone's already answered this and I
> can't find it.
>
> Adrienne
>
> --
> Adrienne Wootten
> Graduate Research Assistant
> State Climate Office of North Carolina
> Department of Marine, Earth and Atmospheric Sciences
> North Carolina State University
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.