Hi !

I have several temporal series where I would like to find peaks. I mean these series are a bit noisy sometimes, with many local maximum but most of them are "small" compared to some and are not interesting to me.

I had this idea, to find peak I am interesting in :

Let Y = y1,y2,...,ynn my temporal serie

I am doing a running average on h points before i [y(i-h),...,y(i)] = Avg(i), for i = h+1 to the length of the serie.
I am doing the same for the standard deviation.

Then, if y(i) - Avg(i-1) > A*std(i-1) with A a factor i need to choose properly but I do not really know yet

If this inequality is verified, I guess i have a max local, not too small;

please, give me some critics or idea if i am doing wrong,

Thank you