Hi,
Consider the line below: for(r in a)for (s in a) x=rbind(x,apply(replicate(1000,V(r,s)),1,mean)) V is a vector of (n-1) variables calculated by some rule and is a functions of (r,s). So the line above produces 1000 replicates of V for each (r,s), puts them in a matrix, calculates the mean of them, and finally puts the means for all (r,s) in a matrix. So the produced matrix, x, is the mean of (n-1) V 's for each possible value of (r,s) in each row. Now for simplicity fix (r,s) in just one point and let n=5. So in each replicate we have only one V which is a vector consisted of 4 variables. Name the elements of V as U1, U2, U3 and U4. Then we can let V(i) = [U1i , U2i , U3i , U4i] which shows each row of V produced per replicate (i=1,2,...,1000). Therefore we can say x=[x1 , x2 , x3 , x4] which is the vector of means calculated at the end. Now what I need is to first calculate the vector below per replicate (i=1,...,1000): Er(i) = [ |Ui1-x1| , |Ui2-x2| , |Ui3-x3| , |Ui4-x4| ] where |A| shows the absolute value of A. Then I should calculate mean of Er(i) 's and put the result in a vector. I just don't know how I can calculate Er(i) 's in the given line above. On the other words, I don't know where I should add the required code in the given line. Thanks for any help in advance! [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see 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.