[CHANGESET]: Statistics function incorrectly computing median

[Ben Abbott]

On Thursday, March 06, 2008, at 09:16AM, "Bateman David-ADB014" <David.Bateman at motorola.com> wrote:
>Why can't 
>
>        for q = 1:n
>          inv(k(q)) = interp1 (cdf, x, p(k(q)), "linear", "extrap");
>        endfor
>
>in __quantile__.m be written as
>
>inv(k) = interp1 (cdf, x, p(k), "linear", "extrap");
>

The original author of the  discrete_inv used to loop over "q" in order to favor choose memory over time (__quantile__ evolved from  discrete_inv).

Not knowing how much data might be encountered, I had left it as it was

However, that said, I'm in favor of changing it (memory is cheap and getting cheaper all the time).

Ben