On Mon, Jun 25, 2012 at 12:46 AM, Ian Clark <earthspo...@gmail.com> wrote:
> Will try out your C function in my context, and try to tweak it. Your
> solution detects when the step happens, solving 1 in my stated
> objectives: 1-4. I fear however that I am more interested in 2-4.

If t=: 1+X %. <:/~i.#X

(which probably does not approximate T in real time, unless X is
relatively short, but you just now stated that that was not your
primary concern)

then these seem like reasonable estimates of the means+K:

m1=: (+/%#) t {. X
m2=: (+/%#) t }. X

Picking an epsilon, for determining if they are equal, or not, would
seem to require some knowledge of the magnitude of the noise (G and V)
and of the possible values for the means.  But I have not seen
anything like that yet in the problem statement.

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to