"mystic" apply ?

hi,

have anybody a suggestion why this function
works only for a vector correct.
When i'm using apply(data,2,fuzzy) i have the columns
stacked (i.e. dim(x) =1000,4  should after the function 1000,8
instead is 2000,4) - for single vector it's ok.

No, apply doesn't do that -- it doesn't look at matrix arguments at all.

If dim(x) is c(1000,4) then fuzzy is called 4 times, each time with an
argument of length 1000.  It returns a value of length 2000 and so you get
a 2000x4 result.

I can't see why you would expect to get a 1000x8 result -- I might have
expected a 1000x4x2 array, though.

You can get the result you want with

matrix(apply(x,2,fuzzy), nrow=nrow(x))


	-thomas

fuzzy  <- function (x)
{
    fuz.x <- cbind(x, x)
    min <- quantile(x, 0.20)
    max <- quantile(x, 0.80)
    fuz.x[x >= max, 1] <- 1
    fuz.x[x >= max, 2] <- 0
    fuz.x[x <= min, 1] <- 0
    fuz.x[x <= min, 2] <- 1
    fuz.x[x > min & x < max, 1] <- (x[x > min & x < max] - min)/(max - min)
    fuz.x[x > min & x < max, 2] <- (max - x[x > min & x < max])/(max - min)
    fuz.x
}

______________________________________________
R-help at stat.math.ethz.ch mailing list
http://www.stat.math.ethz.ch/mailman/listinfo/r-help

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle
^^^^^^^^^^^^^^^^^^^^^^^^
- NOTE NEW EMAIL ADDRESS

many thanks for the enlightenment !

I can't see why you would expect to get a 1000x8 result -- I might have
expected a 1000x4x2 array, though.

I'm working with fuzzy-logic membership
functions for data.frames with real values.
i.e.   age
         23
could get  age.young     age.old
                    0.75            0.25
How many coulmns you can get is a point of definition.
Possible are  modifiers as "small, middle, big"  or many others , but the
sum is always 1.

christian