R help - Mar 2011 - How would you avoid loops in this simple example?

I'm trying to code more R-like and avoid loops. Here is an example I'm
having
a hard time getting away from loops with, and the real matrices are rather
large, and the computation is too time-consuming.

### Dimensions
N <- 2
M <- 3
P <- 4
### Array and Matrices
nu <- array(NA,dim=c(N,M,P))
Lambda <- matrix(1:12,P,M)
F <- matrix(1:8,N,P)
### Loop to avoid
for (i in 1:N) {for (m in 1:M) {for (p in 1:P) {
     nu[i,m,p] <- Lambda[p,m] * F[i,p]
}}}

Any help is appreciated. Thanks.

--
View this message in context:
http://r.789695.n4.nabble.com/How-would-you-avoid-loops-in-this-simple-example-tp3390017p3390017.html
Sent from the R help mailing list archive at Nabble.com.

Hi,

You are basically just doing matrix multiplication, which R has some
built in, optimized functions for.  You can see the documentation at
?matmult
This gets you part of the way there.  I suspect you can even avoid the
last loop if you know enough about linear algebra and how to use some
clever transposition and inner or outer multiplication (?outer ?t
?matmult).

## alternate nu
nu2 <- array(NA,dim=c(N,M,P))
## avoiding much of that loop
for(i in 1:P) {
  nu2[,, i] <- F[, i, drop = FALSE] %*% Lambda[i, ]
}

HTH,

Josh

On Sat, Mar 19, 2011 at 10:27 AM, zerfetzen
<paul.heinrich.dietrich at gmail.com> wrote:
> I'm trying to code more R-like and avoid loops. Here is an example
I'm having
> a hard time getting away from loops with, and the real matrices are rather
> large, and the computation is too time-consuming.
>
> ### Dimensions
> N <- 2
> M <- 3
> P <- 4
> ### Array and Matrices
> nu <- array(NA,dim=c(N,M,P))
> Lambda <- matrix(1:12,P,M)
> F <- matrix(1:8,N,P)
> ### Loop to avoid
> for (i in 1:N) {for (m in 1:M) {for (p in 1:P) {
> ? ? nu[i,m,p] <- Lambda[p,m] * F[i,p]
> }}}
>
> Any help is appreciated. Thanks.
>
> --
> View this message in context:
http://r.789695.n4.nabble.com/How-would-you-avoid-loops-in-this-simple-example-tp3390017p3390017.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.
>


-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/

On Sat, Mar 19, 2011 at 10:27:11AM -0700, zerfetzen
wrote:
> I'm trying to code more R-like and avoid loops. Here is an example
I'm having
> a hard time getting away from loops with, and the real matrices are rather
> large, and the computation is too time-consuming.
> 
> ### Dimensions
> N <- 2
> M <- 3
> P <- 4
> ### Array and Matrices
> nu <- array(NA,dim=c(N,M,P))
> Lambda <- matrix(1:12,P,M)
> F <- matrix(1:8,N,P)
> ### Loop to avoid
> for (i in 1:N) {for (m in 1:M) {for (p in 1:P) {
>      nu[i,m,p] <- Lambda[p,m] * F[i,p]
> }}}
Try the following

  Lambda1 <- aperm(array(Lambda, dim=c(P, M, N)), perm=c(3, 2, 1))
  F1 <- aperm(array(F, dim=c(N, P, M)), perm=c(1, 3, 2))
  nu1 <- Lambda1 * F1
  identical(nu, nu1)
  [1] TRUE

The function array() recycles its first argument, if the 
required dimension of the output is larger. The function
aperm() then puts the indices to the required order.

Hope this helps.

Petr Savicky.

Thanks for the help, and now I have some functions to learn. Both suggestions
work great, really appreciate it.

--
View this message in context:
http://r.789695.n4.nabble.com/How-would-you-avoid-loops-in-this-simple-example-tp3390017p3390549.html
Sent from the R help mailing list archive at Nabble.com.