identify "likely a maximum" point

M

Mika

Hi math guys/gals,

I have an "oscillator", points go up and down, (amplitud or frquency
not constants) and need to identify if the newest point is probably a
maximun or a minimun of such a function.

I say probably, because it doesn't need to be perfect, but in hindsight
you can see that your solution got close to identify most of the
max/min region of that oscillator....

So your solution is doing this:

- the latest point is most probably (or close to) a max/min based on
the recent points history.

Thanks for your time.
Mika
 
J

Jason Lepack

What is your question? If you're asking one, then what does your data
look like?

When you make a post you should pretend that you know nothing about
what you're trying to do and see if you can tell what the post is
talking about.

Cheers,
Jason Lepack
 
M

Mika

Oppps, pushed enter wrong time..

FYI hope this is clearer...

here is the thing, let´s say column A has a list of 500 values
(discrete function). Is there a way to identify max and min relatives
in that list ? (of course, I´m not interested in max and min
absolutes!!!) . I guess some type of criteria need to be done. The
result could be something like:


Colum A, Column B,,,,,,


A1
A2 I'm a max rel.
..
A40 I´m a max rel.
..
..
A105 I'm a min rel.
..
..
A500
 
J

Jason Lepack

Under what conditions would a value be determined a "Max Relative" or
"Min Relative"?

Is A2 a max relative because it's bigger than A1? and is A40 a max
relative because it's bigger than everything in A1 to A39?

Cheers,
Jason Lepack
 
M

Mika

Jason,

because is bigger/smaller than its neighbours.... problem is that you
know the past points but not the future ones of course, that is why can
not be perfect... so in some way has to make an educated guess, like
"in the past x points hasn't be that high/low then probably is a
max/min..."

Again, is not need to perfectly pin point the max or min but must be
close to them. Also don't need to catc 100% but 95% should be
perfect...

I do think this is more math than excel stuff and is related, I guess,
to identify extreme conditions...

cheers
Mika
 
G

Gary''s Student

Based upon history, measure the frequency (period) of the oscillator. If the
intervals of measurement are equal, then you can predict where the max and
mins will occur (you don't even have to look at the data).

Of course as time goes on you should re-estimate the period in case of decay
or drift.
 
Top