try this:
trim <- function(x,prop=.05) { trimlow <-quantile(x,prob=(prop)) trimhigh <- quantile(x,prob=(1 - prop)) x[(x >= trimlow) & (x <= trimhigh)] } On 3/20/07, Jonathan Morse <[EMAIL PROTECTED]> wrote: > > Thanks for the rapid response. > > The inequality with the quantile trim works great, however I am in looking > for a two tailed trim. This could work if I could unite two series in a > command similar to the following: > > trim <- function(x,prop=.05) { > trimlow <- x[x < quantile(x,prob=(1-prop))] > trimhigh <- x[x > quantile(x,prob=(prop))] > trimx <- intersect(trimlow, trimhigh) > return(trimx) > } > > However the intersect function will not return repeated values, which is a > problem. Any thoughts?? > > Thanks. > > Petr Klasterecky <[EMAIL PROTECTED]> wrote: If it is precise > enough for you, you can use > > trim <- function(x,prop=.05) { > trimx <- x[x < quantile(x,prob=(1-prop))] > return(trimx) > } > > Petr > > Jonathan Morse napsal(a): > > Hi, > > > > I am trying to restrict a data set so as not to included > outliers. Specifically, I would like to specify a percentage where a > fraction of observations are eliminated from the data set, much in the same > way that the trimmed mean function works - but leaving the restricted data > set intact. > > > > I have been using a function which will restrict the data set using: > >> trim=function(x,p){ > >> o=order(x) > >> xo=x[o] > >> n=length(xo) > >> tl=round(n*p) > >> print(xo[(tl+1):(n-tl)])} > > > > However I was wondering if anyone knew a more elegant and simple method > to get the same result. > > > > Thanks in advance. > > > > > > > > --------------------------------- > > Don't pick lemons. > > > > [[alternative HTML version deleted]] > > > > ______________________________________________ > > R-help@stat.math.ethz.ch 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. > > > > -- > Petr Klasterecky > Dept. of Probability and Statistics > Charles University in Prague > Czech Republic > > > > --------------------------------- > TV dinner still cooling? > Check out "Tonight's Picks" on Yahoo! TV. > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@stat.math.ethz.ch 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. > -- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem you are trying to solve? [[alternative HTML version deleted]] ______________________________________________ R-help@stat.math.ethz.ch 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.