On Aug 10, 2010, at 11:14 AM, RCulloch wrote:
Hi David,
Thanks very much for that reply! I might be a touch out of my
comfort zone,
but I can see how the loop script works and where I went wrong, but
I'm not
sure if I am asking the correct questions here, or perhaps more
accurately
I'm using the wrong command for the task in question - and as you
say more
info would be better! So.....
I want to split the data by day to look at the proportion of time an
individual spent in each of the eight behaviours - there are 30 rows
(i.e.
individuals).
So I'm going over old code trying to make it better (not that it
could be
worse!), especially trying to make it more efficient!
So my old code did this (manually for each day):
##DAY1##
DAY1 <-cbind($X1.1, scananal$X2.1, scananal$X3.1, scananal$X4.1,
scananal$X5.1, scananal$X6.1, scananal$X7.1, scananal$X8.1)
head(DAY1)
From our earlier efforts:
> DAY[[1]]
[1] 1 2 3 4 5 6 7 8
If you were using the earlier results, then this could be made easier
and more generalizable to an iterative process by:
DAY1cols <- scananal[ , DAY[[1]] ] or just ...
DAY1cols <- scananal[ DAY[[1]] ]
You could replace the 1 by an i in a for loop, or create a list of
"subsets" with lapply.
which would give, for example,
head(DAY1)
[,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
[1,] 14 0 2 1 2 2 0 3
[2,] 23 0 0 1 0 0 0 0
[3,] 0 0 0 0 0 0 0 0
[4,] 0 0 0 0 0 0 0 0
[5,] 0 0 0 0 0 0 0 0
[6,] 0 0 0 0 0 0 0 0
I'd run the following script to get the proportions then bind that
together
with other data
###~~~~DAY1~~~###
## CALC NSCANS PER ID ##
n <- rowSums(DAY1)
## GIVE THE DAY NUMBER TO THE DATAFILE
DAY <- rep(1,30)
## CALC PROPORTION OF TIME IN EACH ACTIVITY ##
scansprop <- as.data.frame(prop.table(DAY1,1))
head(scansprop)
##CALC AS ARC_SINE_TRANSFORMED###
transscan<-asin(scansprop)
head (transscan)
##gives column headings##
names(transscan)
##CHECK IT ALL ADDS TO ONE!! ##
rowSums(scansprop)
##MERGES ALL THE DATA FOR THE DAY
DAY1_SUM <- cbind(n,DAY,DAY1,scansprop,transscan)
Then I would merge each of the days, so this script works, but I
know it is
rather a poor effort in R script to say the least!
That may work perfectly well. I don't know. I didn't follow the logic
because your reference to "proportions" was not made with an
unambiguous definition of a denominator in several locations.
I'm trying to work through this myself, but hit a hurdle in the first
instance!
Not sure if this is any clearer?
Perhaps. Two options: Try to put that sequence into a function that
will return a structure. Then test it. Then lapply it.
Or put it into sequence that makes DAY1_SUM inside a for-loop, take
the results of for() {} on DAY[[i]] and have it assign values to a
structure like a list or a matrix.
--
David Winsemius, MD
West Hartford, CT
______________________________________________
R-help@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.
