Timeline for Finding local maxima/minima with Numpy in a 1D numpy array

8 events

when toggle format	what		by	license	comment
Jun 23, 2021 at 23:40	comment	added	Reed Espinosa		@ankostis Note that simply removing the `+1` and substituting `np.gradient()` for `np.diff` in the code above produces the indices of each minima/maxima as well as their lowest/highest neighbors.
Oct 27, 2017 at 16:27	review	Suggested edits
Oct 27, 2017 at 17:39
May 31, 2016 at 18:15	history	edited	Cleb	CC BY-SA 3.0	removed fluff, highlighting
Jun 23, 2015 at 17:18	comment	added	marcman		I know this thread is years old, but it's worth adding that if your curve is too noisy, you can always try low-pass filtering first for smoothing. For me at least, most of my local max/min uses are for global max/min within some local area (e,g, the big peaks and valleys, not every variation in the data)
Jan 30, 2015 at 13:41	comment	added	ankostis		To avoid this `+1` instead of `np.diff()` use `np.gradient()`.
Mar 24, 2013 at 21:06	comment	added	tktk		This will also act weird if there are repetitive values. e.g. if you take the array `[1, 2, 2, 3, 3, 3, 2, 2, 1]`, the local maxima is obviously somewhere between the 3's in the middle. But if you run the functions you provided you get maximas at indices 2,6 and minimas at indices 1,3,5,7, which to me doesn't make much sense.
Feb 28, 2013 at 17:09	comment	added	danodonovan		nice use of nested numpy functions! but note that this does miss maxima at either end of the array :)
Mar 12, 2012 at 12:35	history	answered	R. C.	CC BY-SA 3.0