[R] interaction of shingles and tapply()

[Roger Levy]

I'm interested in a version of tapply() that operates with shingles 
instead of factors.  For instance:

x <- c(1,1,2,2,3,3)
y <- c(1,1,1,0,0,0)

s <- shingle(x,intervals=cbind(c(0.5,1.5),c(2.5,3.5)))


# the following function should exist!
tapply.shingle(x,s,mean)  # returns the vector c(0.75,0.25)


I've written such a function as follows:


tapply.shingle <- function(x,s,fn,...) {
   result <- c()
   for(l in levels(s)) {
     x1 <- x[s > l[1] & s < l[2]]
     result <- c(result, fn(x1,...))
   }
   result
}


However, I'm not thrilled with the for() loop, and I don't see any way 
to generalize this function to handle lists of shingles instead of 
individual shingles, except to use recursion.  Does anyone have any 
suggestions or thoughts?

Many thanks,

Roger