aggregating a vector

[Francesco Potorti`]

Hi, 

  I often have this problem: given a vector A, compute the variance of
the elements of an aggregated vector obtained by summing up the
elements of the given vector k by k (suppose k is a divisor of
length(A)).  Currently I use the following code:

    do_fortran_indexing = "true";
    sets = length(A)/k;
    YY = zeros (k, sets);
    YY(:) = A(:);
    Y = sum (YY);
    var = (sumsq(Y) - sum(Y)^2/sets) / (sets-1);

Is there anything faster?
--
         Francesco Potorti`     | address@hidden (Internet)
         researcher at          | 39369::pot (DECnet)
         CNUCE-CNR, Pisa, Italy | +39-50-593203 (voice) 904052 (fax)